(navigation image)
Home American Libraries | Canadian Libraries | Universal Library | Community Texts | Project Gutenberg | Children's Library | Biodiversity Heritage Library | Additional Collections
Search: Advanced Search
Anonymous User (login or join us)
Upload
See other formats

Full text of "Graphic presentation simplified"

by R.R.Lutz 







^ 

A guide fo Ike use of 
graphic methods 
in business 






MODERN INDUSTRY IOOKS 



GRAPHIC 

PRESENTATION 

SIMPLIFIED 

ByR. R. LUTZ 

A MODERN INDUSTRY BOOK 

A practical guide to the use of modern graphic 
methods in business, this new book shows you 
how to present facts in their most simple visual 
form . . . how to make all kinds of graphic 
charts . . . how to use them . . . how to analyze 
a problem, select the equipment and execute 
the graph or graphs best suited to the problem. 

Detailed information is given on: 

Curve charts Bar charts Circle 
charts Sector charts Statistical 
map charts Dot maps Pin and 
tack maps Flow maps the use of 
color. 

Invaluable for businessmen, managers and ex- 
ecutives, as well as clerical workers and statis- 
ticians, this book will enable you to analyze 
and record the operations of your business. 

In addition, the book offers a training program 
for beginners who must depend on their own 
efforts without skilled supervision, with clear 
instructions in layout and plotting, and exam- 
ples for practice. 

FUNK & WAGNALLS COMPANY 
153 East 24th Street New York 10. N. Y. 



From the collection of the 



n 



m 



o Prelinger 

i a 

AJibrary 

t P 



San Francisco, California 
2006 



GRAPHIC PRESENTATION SIMPLIFIED 



MODERN INDUSTRY BOOKS 



Your Public Relations GLENN GRISWOLD AND DENNY GRISWOLD 

Standard Business-Conference Technique CARL HEYEL 

How To Reduce Distribution Costs RICHARD D. CRISP 

Practical Handbook of Industrial Traffic Management RICHARD c. COLTON 

Management Controls for Foremen and Supervisors HARRY j. MCCAULLY, JR. 

The Law of Free Enterprise LEE LOEVINCER 

Bargaining with Organized Labor RICHARD c. SMYTH AND MATTHEW j. MURPHY 

Supervision in Business and Industry ROBERT D. LOKEN AND EARL p. STRONG 

Wage Policy for Management SUMNER D. CHARM 

Our New National Labor Policy FRED A. HARTLEY, JR. 

Tested Techniques in Labor A rbit ration GEORGE w. TORRENCE 

The Economics of Industrial Management WALTER RAUTENSTRAUCH AND RAYMOND VILLERS 

Foundations for Constructive Industrial Relations R. CARTER NYMAN 

Management Men and Their Methods LUIS j. A. VILLALON 

Graphic Presentation Simplified R. R. LUTZ 

How To Cut Production Costs H. E. BLANK, JR. 

Teamwork in Industry WILLIAM SEWARD 

How To Sell to Latin America ABRAM A. PRECIADO 




BY R. R. LUTZ 



FUNK & WAGNALLS COMPANY NEW YORK 

m association with 
MODERN INDUSTRY MAGAZINE NEW YORK 



COPYRIGHT, 1949, BY 
MAGAZINES OF INDUSTRY, INC. 

1 



COPYRIGHT UNDER THE ARTICLES OF THE COPYRIGHT CONVENTION 

OF THE PAN AMERICAN REPUBLICS AND THE UNITED STATES 

PRINTED IN THE UNITED STATES OF AMERICA 



ACKNOWLEDGMENTS 



THE AUTHOR wishes to express his deep appreciation for the help 
received in the preparation of this volume from fellow workers with 
whom he has been associated in the course of nearly twenty years' 
experience in graphic design and presentation. He wishes also to thank 
the business firms and associations which have made available for in- 
clusion in the book many of the graphic examples used as illustrations 
in the discussions of methods and principles. He is especially indebted 
to Mr. Willard C. Brinton for his helpful advice and counsel during 
the planning stage of the volume. Grateful thanks are due Mrs. Mabel 
Howell for her painstaking aid in the preparation of the charts and 
other illustrations for publication. 



CONTENTS 



ACKNOWLEDGMENTS v 

LIST OF ILLUSTRATIONS xi 

LIST OF TABLES xix 

CHAPTER PAGE 

I AN EXPANDING ART 1 

Graphic Presentation in Two World Wars Charts 
Versus Tables Improvements in Equipment Re- 
duction in Length of Training Training Methods 

II CURVE CHARTS 10 

Designation of Chart Components Titles Abbrevi- 
ations for Time Scales Proportions of Width and 
Height Scale Numbers Position and Labeling of 
Scales Scale Rulings Curve Patterns Omission of 
Zero Base 100 Percent Base in Index-number Charts 
Break in Time Scale Fitting Curves to Gain Space 
Omission of Time-scale Ruling Above Curves- 
Plotting Points Moving-average Curves Weight of 
Grid Lines Broken Curve Stock-price Charts Cu- 
mulative-curve Charts Multiple-scale Charts Mul- 
tiple-time-scale Charts Staircase-curve Charts Keys 

vii 



viii CONTENTS 

CHAPTER PAGE 

III SURFACE-CURVE CHARTS 32 

Differentiation of Trends by Shading Strata Charts 
Distortion Produced by Steep Curves Use of Shad- 
ing to Show Net Results Shading Significant Time 
Periods Contrasting Net Changes Miniature Sur- 
face-curve Charts Long-range Business - activity 
Charts Shaded Staircase-curve Charts Shading 
Versus Solid Black Economy of Space in Design 

IV RATE-OF-CHANGE CHARTS 44 

Grid Ruling Omission of Scale Rulings Rate-of- 
change Versus Index Numbers Methods for Laying 
Out Grids Explanatory Note Units Varying in Size 
and Kind Limited Use of Rate-of-change Designs 

V BAR CHARTS 54 

Wide Range of Values Lettering Versus Scales- 
Order of Components and Ruling Lettering Stub on 
Bars Joined-bar Charts Grouped-bar Charts Keys 
Paired-bar Charts Bar-and-symbol Charts Area- 
bar Charts Optical Distortion in Bars Inclusion 
or Omission of 100 Percent Base Deviation-bar 
Charts Base Figures in Charts of Relative Numbers 
Adjustment of Stub to Space Step-bar Charts 
Sliding-bar Charts 

VI COLUMN CHARTS 76 

Column and Segment Labels Comparison of Col- 
umns with Curves Hollow Columns Combination 
of Spaced and Joined Columns Combination of 
Chart and Table Grouped-column Charts Com- 
bination of Columns and Curves Charts Showing 
Both Basic Data and Percentage Distribution Devi- . 
ation-column charts 

VII CIRCLE CHARTS 90 

Common Errors in Drafting Circles Short-cut Meth- 
od for Calculating the Size of Circles Circles Com- 
pared with Bars 



CONTENTS ix 

CHAPTER PAGE 

VIII SECTOR CHARTS 96 

Order of Sectors Lettering of Sector Charts Group- 
ing of Sectors Miniature Sector Charts Separation 
of Sectors for Emphasis Omission of Shading Sec- 
tor Totals 

IX STATISTICAL MAP CHARTS 104 

Use of Colors Maps with Surface Designations of 
Categories Dot Maps Maps with Circles Maps 
with Columns, Bars, or Curves Pin and Tack Maps 
Flow Maps Keys 

X PICTORIAL CHARTS 112 

Bar Pictorial Charts 

XI MISCELLANEOUS TYPES 116 

Ranking Charts Frequency - distribution Charts 
Relative-variation Charts Organization Charts 
Comparison of Equal Amounts with Circles Circu- 
lar-bar Design 

XII ELEMENTARY TRAINING METHODS 127 

Charts on Printed Graph Paper Typewritten Charts 

XIII SMALL CHARTS ON UNRULED PAPER 132 

Equipment Grid Lay-out Types of Scales Joined- 
bar Lay-out Spaced-bar Lay-out Column-chart 
Lay-out Curve-chart Lay-out and Plotting Circle- 
and Sector-chart Lay-out Lettering Paper Inking 
Order of Operations 

XIV DISPLAY CHARTS 146 

Use of Color Lay-out and Equipment Lettering- 
Paints Crayons Airbrush Coloring Rolled Charts 
Mounting on Cardboard 

XV REPRODUCTION PROCESSES 153 

Photostat-Photo-offset-Multilith-Stencil and Gela- 
tin Duplicators Enlargements Reductions 



x CONTENTS 

CHAPTER PAGE 

XVI CLASSIFICATION OF CHARTS BY USE 160 

Charts for Publication Charts for Inter-office Distri- 
butionCharts for Desk Use Charts for Conferences, 
Board Meetings, and Hearings Chart Exhibits 
Charts for Public Addresses Charts for Recording 
Current Operations 

XVII STATISTICAL PROCEDURES 175 

Rounding Averages Weighted Average Frequency 
Distributions Index Numbers 

XVIII PLANNING, TABULATION, AND PRESENTATION 

OF TABLES 189 

Tabulating Sheets-Units of Value-Titles-Stubs- 
TabulationColumn Widths Capitalization Punc- 
tuation Symbols Ruling 

INDEX 197 



LIST OF ILLUSTRATIONS 



CHARTS 

CHART PAGE 

1 Reduction in Schedules of Production, Combat Planes: 

December, 1917, through February, 1918 2 

2 Balance of Construction and Sinkings, British Merchant 

Marine: 1914 to 1916 4 

3 Designation of Chart Components 11 

4 Effects of Scale Alterations 12 

5 Too Much Ruling 16 

6 Ruling Replaced by Ticks 16 

7 Combination of Ticks and Ruling 16 

8 Curve Patterns 17 

9 Omission of Zero Base 19 

10 Wages, Prices, and Production 20 

11 Retail Cost of Food: 1939, 1944, and 1945 21 

12 Depression, Recovery, Depression 23 

13 Stocks of C, D, and K Emergency Rations: 1943 24 

14 Plotting to Space or Line 25 

15 Business Failures 26 

16 Placements by U.S. Employment Service: 1937 to 1945 27 

xi 



xii LIST OF ILLUSTRATIONS 

CHART PAGE 

17 Broken Curve 28 

18 Stock-price Chart 29 

19 Cumulative-curve Chart 29 

20 Multiple-time-scale Chart 30 

21 War Appropriations: 1917-18 and 1940-41 31 

22 Depression and Recovery 33 

23 Production of Motor Trucks and Truck Tractors: 1936 to 

1945 34 

24 Distortion Produced by Steep Curves Rising or Falling at 

a Uniform Rate 35 

25 Average Weekly Earnings of Wage-earners in Manufac- 

turing, and Consumers' Price Index: 1939 to 1945 36 

26 Money in Circulation, United States, Per Capita: 1916 to 

1945 37 

27 Immigrant Aliens Admitted and Departed: 1932 to 1945 38 

28 Tonnage Received and Shipped by QMC Depots: 1944 39 

29 American Business Activity 40 

30 Employment and Unemployment, United States: 1929-1939 41 

31 Hours Spent in the Air by American Service Planes at the 

French Front During World War I 42 

32 Volume of Production, United States: 1859 to 1934 43 

33 Arithmetic and Rate-of-change Scales 45 

34 Top and Bottom of Cycle Omitted 46 

35 Growth of Population, Puerto Rico and Hawaii: 1930 to 

1940 47 

36 Method for Laying Out Rate-of-change Grids with Log- 

arithmic Paper 48 

37 Scale for Laying Out Rate-of-change Grids 49 



LIST OF ILLUSTRATIONS xiii 

CHART PAGE 

38 Industrial Contributions of Science and Invention, United 

States: 1830 to 1930 51 

39 Arithmetic, Logarithmic, and Index Log Charts 52 

40 Receipts of Wheat at Primary Markets: 1943 55 

41 Crime Rates in Urban Communities, by Geographic Divi- 

sions: 1943 (Data Scaled) 56 

42 Crime Rates in Urban Communities, by Geographic Divi- 

sions: 1943 (Data Lettered on Bars) 57 

43 Lettering Within and At the End of Bars 58 

44 Post, Camp, and Station Stocks of Non-perishable Subsis- 

tence Items 59 

45 Federal Expenditures Per Capita 60 

46 Population of Five Largest States: 1940 61 

47 Index of Stocks, Secondary Items 62 

48 Two Methods for Placement of Keys 63 

49 Net Income or Deficit of Governmentally Owned or Op- 

erated Railways in Various Foreign Countries: 1935 or 

1936 64 

50 Contractor-owned Termination Inventory 65 

51 Proportion of Working Population Covered by the Old Age 

Provisions of the Social Security Act, Using the 'Dis- 
tribution of Occupations of the 1930 Census 66 

52 Optical Distortions 67 

53 Inclusion or Omission of 100 Percent Base 68 

54 Price Changes 69 

55 Army Funds Uncommitted 70 

56 Consumers' Prices 71 

57 Estimated Non-agricultural Employment, Five Largest In- 

dustrial Groups: Last Quarter, 1945 72 



xiv LIST OF ILLUSTRATIONS 

CHART PAGE 

58 Exports (Including Reexports), United States, by Conti- 

nents: 1940 73 

59 Loans and Investments of the Member Banks of the Fed- 

eral Reserve System from 1929 to 1935 74 

60 Disposition of Surplus Property: June to December, 1944 77 

61 Instalment Credit: 1939 to 1945 (Column Design) 78 

62 Instalment Credit: 1939 to 1945 (Curve Design) 79 

63 Men Sailing Each Month to France and Home: 1917 to 1919 80 

64 Industrial Production, Manufactures: 1939 to 1945 81 

65 Industrial Production 82 

66 Heavy and Light Bombers Accepted: 1942 83 

67 Par and Market Values of Bonds Sold on New York Stock 

Exchange: 1938 to 1945 84 

68 Receipts, Expenditures, and National Debt 85 

69 Method for Lettering Basic and Percentage Figures on Col- 

umn Charts 86 

70 Interest Payments on Debt of Federal, State, and Local 

Governments: 1932 and 1945 87 

71 Total Exports and Imports of Merchandise: 1915, 1925, 

1935, and 1945 88 

72 Domestic Corporate Security Issues 89 

73 Use of Different Geometric Forms in Comparing Sizes 91 

74 New Industries and Employment (Circle Design) 94 

75 New Industries and Employment (Bar Design) 95 

76 Comparison of Sector and Bar Charts 97 

77 Imports, United States, by Economic Classes: 1942 98 

78 Contract Awards to Small Firms: 1944 99 

79 Composition of National Guard Divisions, World War I 100 



LIST OP ILLUSTRATIONS xv 

CHART PAGE 

80 Causes of Delays in May Deliveries, Signal Corps: 1942 101 

81 Distribution of Male Labor Force 14 Years Old and Over, 

United States: 1940 102 

82 Military Expenditures, United States: 1939 103 

83 Percent of Drafted Men Passing Physical Examination, by 

States, World War I 105 

84 Number of Contracts Awarded in Each State in 1944 106 

85 Rural Rehabilitation Cases Receiving Advances of Capital 

or Goods: 1935 107 

86 Distribution of Florida Grapefruit in the 1932-33 Season 109 

87 Typical Distribution of Employees in a Large Telephone 

Company 113 

88 Printed Symbols for Use in Bar Pictorial Charts 114 

89 Outlook for Surplus Property Disposals 115 

90 Rank of Commodity Groups Based On Wholesale Prices: 

1939 Average and December, 1945 117 

91 Rank of Each of the United States in Ten Educational Fea- 

tures: 1910 118 

92 Estimated Population, United States, by Age Groups: 1945 119 

93 Estimated Weekly Hours Worked, Non-agricultural Em- 

ployment: Third Quarter, 1945 120 

94 Relative Monthly Variation Stocks, Selected Equipment 

Items: 1944 121 

95 Relative Monthly Variation Non-perishable Subsistence 

Placed on Contract, Selected Items: 1944 122 

96 Typical Organization-chart Design 123 

97 The Deficit in Perspective 124 

98 Exports, U.S. Merchandise, as Percent of Total Exportable 

Goods: 1909 to 1945 125 



xvi LIST OF ILLUSTRATIONS 

CHART PAGE 

99 Expenditures for New Private and Public Construction: 

1934 to 1945 128 

100 Employment of Women, Selected Durable Goods Indus- 

tries: 1941 and 1942 

101 Use of Scale in Grid Lay-out 

102 Vertical Distribution of Bars 

103 Method for Plotting Amounts 

104 Bars and Spacing in Spaced-bar Charts 

105 Columns and Spacing in Spaced-column Charts 

106 Plotting of Amounts on Column Charts 

107 Curve-chart Lay-out 

108 Circle- and Sector-chart Lay-out 

109 Example of Stripped Title and Curve Labels 

110 D. C. Building Boom 

111 QMC Military Personnel 

112 Employment and Average Weekly Hours, Wage-earners in 

Manufacturing: 1929 to 1937 182 

113 Average Employment and Hours Worked 

114 Production, Wholesale Price, and Value of Production 

115 Output Per Man and Per Hour 



FIGURES 

FIGURE PAGE 

1 Samples of Adhesive Screens 7 

2 Map Tacks for Large Map Charts 

3 Samples of Graph-paper Ruling 

4 Triangular Decimal Scale 



LIST OF ILLUSTRATIONS xvii 
FIGURE PAGE 

5 Tape Scales 135 

6 Compass and Protractor 142 

7 The Vari-Typer 143 

8 Ruling Pen 144 

9 Bow and Drop Pens 145 

10 Drafting Table Fitted with Parallel Rule 148 

11 Leroy Lettering Guide 149 

12 Sample Lettering-guide Alphabet 150 

13 Method for Stiffening Rolled Charts 152 

14 Comparison of Negative and Positive Photostat Prints 154 

15 Reduction of Lettering and Ruling 156, 157 

16 Reduction of Elite Type 158 

17 Portable Fixture for Exhibit of Small Charts 169 

18 Wall Cross-section of a Chart Exhibit 171 

19 Lantern-slide Projector 172 



LIST OF TABLES 



TABLE PAGE 

I Root-two Dimensions 15 

II Calculation of Circle Diameters 92 

III Square Roots of Decimals-.Ol to .99 93 

IV Simple Average 177 
V Average of a Wide Range of Values 177 

VI Arrangement of Values to Determine the Median 178 

VII Weighted Average 178 

VIII Simple Frequency Distribution 179 

IX Class-interval Distribution 180 

X Differences in Range of Class Intervals 180 

XI Employment and Hours Worked Per Week, Manufacturing 

Industry: 1929 to 1937 182 

XII Employment, Hours Worked Per Week, and Total Hours 

Worked: 1929 to 1937 183 

XIII Production, Wholesale Prices, and Value of Production: 

1929 to 1937 184 

XIV Output Per Man and Per Hour, Manufacturing: 1929 to 

1937 186 



xx LIST OF TABLES 

TABLE PAGE 

XV Acquisition, Disposal, and Inventory of Surplus Property 

through March 31, 1946 190 

XVI Storage of Major Groups of ASF Items: 31 July 1944 192 



GRAPHIC PRESENTATION SIMPLIFIED 



AN EXPANDING ART 




I HE major development and utili- 
zation of graphic presentation as a medium for translating statistics into 
condensed visual form have taken place during the present century. 
Statistical charts were employed previously to a limited extent by 
specialists in various branches of science, but were little used in the 
world of commerce and industry. Much of the progress achieved in 
the last fifty years must be credited to the pioneers who visualized the 
potentialities of a wider employment of graphic presentation in these 
fields. Worthy of special mention is the work of Willard C. Brinton, 
whose Graphic Methods for Presenting Facts was published in 1914. 
In the following decade there appeared a number of excellent treatises 
on this subject, among them Karsten's Charts: How to Make and Use 
Them, and Riggleman's Graphic Methods for Presenting Business 
Statistics. Brinton's encyclopedic Graphic Presentation, containing prac- 
tically every known chart design, was one of the outstanding contribu- 
tions in the decade 1930-40. Time Series Charts, A Manual of 

I 



o 

cr 



x 
o 

2 
LU 

cr 



LU 
_J 
CD 



LU 



"I 
<x 

00 
CD 




10 I I 

5il U 



00 



(M 



0(T> QJ 0)0) 

-3 LL U. U- 

t t tt 






o 

Z) 
Q 
O 

or 
o_ 



CO 
LU 

_l 

ID 
O 
LU 

X 

o 

CO 




o 
o 



o 
o 



o 

o 



00 
I I 
05 



1 







fc K 

? g 



05 



o 

o 



c 



Cvj 



AN EXPANDING ART 3 

Design and Construction, prepared by the Committee on Standards 
for Graphic Presentation under the procedure of the American 
Standards Association, published in 1938, marked an important ad- 
vance in the standardization of nomenclature and methods of pres- 
entation. 

GRAPHIC PRESENTATION IN TWO WORLD WARS 

A comparison of the use of graphic methods by Army agencies 
during the last war with the extent they were employed to record and 
summarize military statistics in World War I gives some indication of 
the advances realized in this field during the last twenty -five years. 

For some time during World War I, the production of statistical charts 
was centered chiefly in the Statistics Branch of the General Staff. As a 
member of that agency, the writer had many opportunities to observe 
the methods employed in the various services for presenting statistical 
data. Of figures there was no end, but few attempts were made to 
reduce them to graphic form. Only a limited number of the higher 
officers were familiar with this type of presentation. 

In the General Staff the value of compact graphic summaries of 
the principal statistical facts of the war soon gained recognition. Early 
in 1918, the Statistics Branch established a system of weekly reports 
for circulation in the General Staff. These reports consisted of sta- 
tistical charts accompanied by text discussion and summaries. This 
service was supplemented by large display charts for presentation 
in weekly conferences attended by the Secretary of War, the Chief 
of Staff, and a selected group of officers representing the various 
services. 

Copies of two charts shown in these conferences appear on pages 2 
and 4. Both have a certain historical interest. Chart 1 played a 
prominent part in the crisis which finally resulted in the removal of 
combat plane and engine procurement from the Signal Corps to an- 
other procurement agency. The curve chart shows the successive reduc- 
tions in the schedules of production from December, 1917, through 
February, 1918. The bars at the right indicate the effect of these reduc- 
tions, after allowing for transportation to Europe and other delays, on 
the projected supply of fighting planes available at the French front 
by the following July. 

Chart 2 pictures the shipping situation resulting from the submarine 
campaign against the British Merchant Marine from the beginning of 



GRAPHIC PRESENTATION SIMPLIFIED 




90 

QUARTERS 3 4 

1914 



CHART 2. BALANCE OF CONSTRUCTION AND SINKINGS, BRITISH MERCHANT 
MARINE: 1914 TO 1916. 



the war in 1914 to the end of 1916. The 100 percent line represents the 
total tonnage at the outbreak of war. The black portions of the curve 
indicate the new tonnage built, the shaded portions the tonnage sunk. 
The end of the curve in each quarter shows the net balance to that date. 

The subjects discussed in these conferences were limited to questions 
of major importance in the conduct of the war. Few of the officers 
attending them were accustomed to the use of charts as a medium for 
the presentation of statistical data. Some of them found rather puzzling 
the conversion of the tabular forms with which they were familiar to 
graphic arrangements of curves, bars, and columns. At the end of the 
first conference one of the officers, a lieutenant general, confessed that 
he had been able to understand only the bar charts and added the com- 
ment that the other types, particularly the curve designs, were too tech- 
nical. 

Experience quickly demonstrated the need for extreme simplicity 
in design and the organization of complicated presentations into 
independent units limited to single aspects of the subject-matter. 
Every effort was directed toward making each chart so clear that its 
meaning and significance could be grasped in the shortest possible 
time. Details which did not bear directly on the purpose of the 
chart were rigorously excluded. Verbal definitions and explanations 



AN EXPANDING ART 5 

were reduced to a minimum by clear labeling of each element in the 
design. 

Within a few months after the declaration of war in 1941, facilities 
for the production of statistical charts were installed by every major 
procurement agency of the Army. Several months earlier, Colonel, later 
General, Leonard P. Ayres, who directed the work of the General Staff 
Statistics Branch in World War I, was called to active service to head 
the central statistical organization. The adoption of graphic methods in 
the various services was furthered by the active support of many 
officers from civil occupations, where their advantages were generally 
recognized. 

Early in the war, arrangements were made by General Ayres for 
monthly meetings with the House and Senate Military Committees for 
the presentation of current information concerning the procurement of 
supplies and military equipment. For this purpose large display charts 
in color were employed. The sessions were short, lasting usually from 
one to two hours. Only through the facility for condensation afforded 
by charts was it possible to cover in so short a time the required range 
of subjects and volume of data. Many knotty questions arising in the 
course of the discussions were quickly cleared up by charts which com- 
pressed on a few square yards of cardboard the pertinent facts from 
tables covering scores of pages. 

CHARTS VERSUS TABLES 

Charts are the shorthand of statistics. In a period in which every kind 
of factual information tends more and more to expression in statistical 
terms, they have become one of the most effective means for converting 
masses of data to a form that facilitates rapid comprehension and inter- 
pretation. Statistical information presented graphically has the great 
advantage over tables that it is more easily understood and remembered 
than the same data in tabular form. This applies particularly to inter- 
related factors, which in the chart form part of an integrated whole, 
while in the table they appear in unconnected detail. The coordination 
of significant relationships made apparent by the graphic method is 
of special value to executives whose duties leave them little time for 
analysis of data presented in tables. 

Charts serve better than any other method of presentation where data 
must be presented to a number of persons at the same time. Statistical 
tables cannot be used effectively in a conference, board meeting, or 



GRAPHIC PRESENTATION SIMPLIFIED 



hearing. The employment of display charts for such purposes has kept 
pace with the increasing use of the conference method for discussion 
and study of economic and legislative problems. 

IMPROVEMENTS IN EQUIPMENT 

Shortening of the training period has been made possible by innova- 
tions in equipment which have reduced the needier long practice and 
a high degree of manual skill. The introduction of two inventions, 
mechanical lettering guides and a new type of adhesive screen for the 
application of shading and color, has greatly simplified the drafting 
techniques employed in the production of statistical charts. Lettering 
of titles, scales, labels, and stubs usually takes from one half to two 
thirds of the time required to complete a chart. Before lettering guides 
came into common use, the proportion of time for this part of the work, 
done by hand, was much greater, while the results were inferior in 
uniformity and appearance to those obtainable with the guides. Free- 
hand lettering is seldom seen now except for ornamental lettering or on 
very large charts where the size of the lettering exceeds the capacity 





25-R 



97 



56 





80-R 





26 



47 



19 





77 



24-R 



20-R 



The Para-Tone Company 

FIGURE 1. SAMPLES OF ADHESIVE SCREENS. 



AN EXPANDING ART 7 

of the guides. Relatively few statistical draftsmen can do good freehand 
lettering. In an extensive check of drafting rooms in Washington the 
average was less than 25 percent. 

The use of adhesive screens for shading and color has cut down ma- 
terially the time, labor, and skill required for filling in surfaces on statis- 
tical charts. Hatching by hand or with a section liner was a tiresome 
and time-consuming job, and the product often lacked uniformity in 
lining and spacing. In addition the patterns of shading that could be 
done by hand were limited in number and variety. A few samples of the 
many designs now available are shown in Figure 1. They are backed 
with a special adhesive which permits removal and replacement, if cor- 
rections or substitutions are necessary, without defacing the surface of 
the chart. They can be applied in about one tenth the time that would 
be needed to duplicate them by hand. 

Formerly, where color was employed, poster paints or water-colors 
were painted on the chart with a brush. The process was extremely slow 
as compared with the speed attainable with adhesive colors, and the 
results were less finished and attractive. The saving in time is around 60 
to 75 percent. 

REDUCTION IN LENGTH OF TRAINING 

Today practically all large enterprises maintain graphic records of 
their internal operations. That the use of these methods has not spread 
to the same extent among the smaller business firms is partly due to the 
common assumption that the preparation of statistical charts requires 
the services of highly skilled draftsmen. The experience gained during 
the last war in the training of drafting apprentices does not bear out this 
assumption. On account of the heavy demands of the military agencies 
and the shortage of draftsmen, it became necessary to recruit drafting 
personnel from every possible source. In Washington alone a large num- 
ber of clerks, typists, and other clerical employees were enlisted for 
apprenticeship courses in drafting. Within a surprisingly short time 
most of them were able to turn out work of acceptable quality. 

Many of these recruits were women, with a sprinkling of young men 
who failed to pass the physical tests for the draft. Practically all ac- 
cepted willingly the opportunity to learn statistical drafting. With some 
the higher rate of pay was the chief incentive. Others welcomed the 
release from monotonous clerical duties offered by the new occupation. 
A creative activity that gives play to imagination, ingenuity, and the 



8 GRAPHIC PRESENTATION SIMPLIFIED 

exercise of manual skill has attractions lacking in the average clerical 
job. 

An experiment conducted by a Federal agency in 1947, involving the 
use of partly trained drafting personnel, showed some interesting re- 
sults. The project was the establishment of a system of control charts in 
each of thirty-three regions throughout the country. Many difficulties 
were encountered. In some localities the services of trained draftsmen 
were unobtainable. In several instances it became necessary to send 
skilled personnel from Washington to get the projects started. The ex- 
perience was enlightening, particularly where, due to the shortage of 
draftsmen, it became necessary, in order to meet the emergency, to 
develop a drafting force recruited from office employees engaged in 
other occupations. On the whole, these makeshift methods were success- 
ful. Within the time limit set, a matter of two months, each of the 
regions was able to show the status of its operations in a series of fairly 
presentable charts. 

TRAINING METHODS 

Apprentice training during World War II was carried on chiefly in 
large drafting rooms under the direction of experienced supervisors and 
chief draftsmen. Beginners were first taught the use of a few drafting 
instruments. Following this, they were assigned simple tasks under close 
supervision on current jobs in process. Through the instruction they re- 
ceived from the chief draftsman and their observation of the* work per- 
formed by other members of the drafting staff, they gradually acquired 
a working knowledge of the use of scales, plotting methods, application 
of shading and color, and lettering with mechanical guides. The time 
required for development of average speed in lining, plotting, and let- 
tering generally ranged from one to three months. 

The conditions were those which usually exist where the volume of 
work is large enough to warrant specialization. The selection of designs 
usually was made by a member of the statistical staff in consultation 
with the supervisor. Details of the design and lay-out were sketched by 
the chief draftsman, who also wrote the instructions concerning the 
dimensions and spacing of the chart, what scales to use, the weight of 
the lines, the curve patterns, the color and shading, and the size and dis- 
tribution of the lettering. The draftsman had only to carry out these 
instructions. Any question or doubt arising in the course of the work was 
quickly resolved by consultation with the chief draftsman. 



AN EXPANDING ART 9 

This method, which closely resembles those employed by industry in 
training beginners for specialized mechanical operations, has the advan- 
tage that it develops a fair degree of productive capacity in the shortest 
possible time. Its weakness, from the standpoint of the embryo drafts- 
man, is that it gives him little opportunity to acquire the general back- 
ground he will need if later he finds himself in a position where he must 
take the responsibility for the entire process, involving not only the 
manual execution of the chart, but the selection of the design and com- 
bination of the graphic elements to bring out clearly the purpose of the 
chart. 

The facilities afforded by ample equipment and skilled supervision 
found in the drafting installations of governmental agencies and large 
corporations are generally lacking in small commercial and industrial 
enterprises. Apprentice training of the type described in the preceding 
paragraphs is out of the question. The beginner can make a start by tak- 
ing a short course in drafting. The long study and practice required for 
mechanical or architectural drafting is unnecessary. However, the aver- 
age teacher of drafting knows little or nothing of graphic presentation. 
At the end of the course, the student will have acquired some profi- 
ciency in the use of drafting instruments, probably sufficient to enable 
him to carry on under conditions where the major part of the thinking 
and planning is done by others, but entirely inadequate if he is faced 
with the problems of selecting and adapting graphic designs to the 
requirements of complicated statistical presentations. 

The method of training suggested in later chapters of this volume 
was planned for those who must depend on their own efforts, without 
the aid of skilled supervision and instruction. It comprises a careful 
reading of Chapters II to XVI, in which are described the various 
graphic types commonly employed in statistical charts, the uses for 
which each is adapted, their limitations, and their relation to the 
many forms in which the results of statistical tabulation, research, 
and analysis are embodied. Included are detailed instructions in 
methods of lay-out and plotting, with examples -for practice on small 
charts with the tools immediately at hand, followed by suggestions 
and rules for planning and drafting charts of greater size and more 
complicated designs, involving the use of additional types of equip- 
ment. The principle is that of learning by doing. It will insure at 
least the acquisition of what so many draftsmen lack, a thorough 
understanding of the why as well as the how of each step of the 
process. 



CURVE CHARTS 




i HE curve chart, sometimes re- 
ferred to as a "line" chart, is one of the oldest forms of graphic pres- 
entation. It is most frequently used in presenting time series, in 
which the data correspond to successive time periods. 

The curve design is generally employed in charts showing projection 
of trends, or where exact plotting for close reading is required. It is not 
well suited for charts of short time series nor where the purpose of the 
chart is to show comparisons of size or amount rather than the move- 
ment and direction of change. 

DESIGNATION OF CHART COMPONENTS 

Chart 3 shows the terms generally employed for designating the vari- 
ous curve-chart components. Many of the terms apply also to other 
graphic types. These designations, with slight modifications, follow the 
terminology approved by the American Standards Association. 

A variation of minor importance is "stub," sometimes used to desig- 

10 



I 

CO 




II 



12 



GRAPHIC PRESENTATION SIMPLIFIED 



ORIGINAL SCALE 
ARRANGEMENT 



JAN 
CONTRACTING TIME SCALE 



APR 



JAN APR 



CONTRACTING TIME AND 
EXPANDING AMOUNT SCALE 




JAN APR 



EXPANDING TIME SCALE 



JAN 



APR 



EXPANDING AMOUNT SCALE 




JAN 



APR 



CONTRACTING AMOUNT SCALE 



JAN 



APR 



CONTRACTING AMOUNT AND 
EXPANDING TIME SCALE 



JAN 



APR 



CHART 4. EFFECTS OF SCALE ALTERATIONS. FROM Time Series Charts, 
A Manual of Design and Construction. 



CURVE CHARTS 13 

nate the abbreviated scale rulings in time and amount scales, called in 
this example "ticks." "Stub" is confusing, for the reason that it applies 
also to the list of items in bar charts. Another term not shown is "grid," 
signifying the total plotting area, bounded by the side, top, and 
bottom lines of the chart. 

TITLES 

The title should be brief, but without sacrifice of clarity or omission of 
essential details. As to position, above or below the grid, the practice 
varies. In charts appearing in periodicals it is often placed below the 
drawing. Titles of display charts are generally lettered above the grid. 
If economy of space is important, as in miniature charts, the titles, if not 
too long and not accompanied by subtitles, may be lettered within the 
grid. If the chart is surrounded by a border, the title is centered between 
the vertical border lines, not over the space occupied by the chart. 

ABBREVIATIONS FOR TIME SCALES 

Time scales are in units of years, months, weeks, or days. Frequently 
in charts of yearly data there will not be room to letter the full figures, 
in which case two digits with an apostrophe may be substituted, as '39, 
'40. If there is enough space, three-letter abbreviations should be used 
for monthly time units. If this results in crowding, single-letter abbre- 
viations may be employed. Time-scale abbreviations are not followed 
by a period. Time units in weeks should be lettered in figures, with the 
month lettered horizontally below. The clearest form is 1-7, 8-14, and 
so on. Time units in days of the week should be designated by three- 
letter abbreviations. 

PROPORTIONS OF WIDTH AND HEIGHT 

The impression created by a curve chart depends to a great extent on 
the shape of the grid and the distribution of time and amount scales. 
Overexpansion or contraction of one of the scales throws the graphic 
picture out of balance and makes interpretation difficult. Chart 4 illus- 
trates the effects of different scale alterations. 

While no fixed rules for the proportions of grid dimensions can be 
laid down, a convenient standard is the method known as "root-two," a 
ratio of 1 (short side) to 1.414 (long side). These proportions approxi- 



14 GRAPHIC PRESENTATION SIMPLIFIED 

mate those of standard sizes of paper and illustration board. The table 
of root-two dimensions on page 15 will save time and calculation of grid 
measurements. If for any reason either dimension is predetermined to 
fit a given space, the other dimension can be found immediately by 
reference to the table. 

These proportions are suggested as a general guide. Modifications 
will often be necessary if the shape of the chart has to be accommo- 
dated to space requirements of other material, as, for example, column 
widths in periodicals, reports, and other printed matter. 

SCALE NUMBERS 

Amount scales should be in even numbers or in multiples of five or 
ten. No definite rule can be laid down concerning the number of 
amount-scale rulings, but ordinarily they should not exceed five. 

POSITION AND LABELING OF SCALES 

The position of the amount scale is to the left of the grid, close to 
and exactly opposite the scale rulings. In wide charts covering a large 
number of plottings the scale may be repeated at the right of the grid. 

The amount-scale unit, if rounded, should be lettered under the title. 
It should specify not only the size of the unit, but what it represents, as 
"Millions of dollars," "Thousands of men," etc. If this is done, the cap- 
tion over the scale can be shortened to a single word, "Millions" or 
"Thousands," a considerable advantage in drafting, as space for scale 
captions is usually limited. 

Scale labels should not be lettered vertically at the left or right of the 
scale column. All lettering should be easily read without turning 
the chart sidewise. 

If monthly series extend over a year, the end of each year should be 
indicated by a heavier vertical rule or by the symbol | between 
December and January. Both methods are shown in 'Chart 3. 

SCALE RULINGS 

The number of scale rulings for time and amount scales depends on 
the nature and distribution of the data and the shape and dimensions 
of the grid. The extremes in practice range from too little to too much. 
Chart 5 shows an example of too much. The multiplicity of rules clut- 



CURVE CHARTS 15 

TABLE I 
ROOT-TWO DIMENSIONS 



Short 


Long 


Short 


Long 


Short 


Long 


Short 


Long 


Short 


Long 


Short 


Long 


Side 


Side 


Side 


Side 


Side 


Side 


Side 


Side 


Side 


Side 


Side 


Side 


1.0 


1.4 


4.5 


6.4 


8.0 


11.3 


11.5 


16.3 


15.0 


21.2 


18.5 


26.1 


1.1 


1.6 


4.6 


6.5 


8.1 


11.5 


11.6 


16.4 


15.1 


21.4 


18.6 


26.3 


1.2 


1.7 


4.7 


6.7 


8.2 


11.6 


11.7 


16.5 


15.2 


21.5 


18.7 


26.5 


1.3 


1.8 


4.8 


6.8 


8.3 


11.7 


11.8 


16.7 


15.3 


21.6 


18.8 


26.6 


1.4 


2.0 


4.9 


6.9 


8.4 


11.9 


11.9 


16.8 


15.4 


21.8 


18.9 


26.7 


1.5 


2.1 


5.0 


7.1 


8.5 


12.0 


12.0 


17.0 


15.5 


21.9 


19.0 


26.9 


1.6 


2.3 


5.1 


7.2 


8.6 


12.2 


12.1 


17.1 


15.6 


22.0 


19.1 


27.0 


1.7 


2.4 


5.2 


7.4 


8.7 


12.3 


12.2 


17.3 


15.7 


22.2 


19.2 


27.1 


1.8 


2.6 


5.3 


7.5 


8.8 


12.4 


12.3 


17.4 


15.8 


22.4 


19.3 


27.3 


1.9 


2.7 


5.4 


7.6 


8.9 


12.6 


12.4 


17.5 


15.9 


22.5 


19.4 


27.5 


2.0 


2.8 


5.5 


7.8 


9.0 


12.7 


12.5 


17.7 


16.0 


22.6 


19.5 


27.6 


2.1 


3.0 


5.6 


7.9 


9.1 


12.9 


12.6 


17.8 


16.1 


22.7 


19.6 


27.7 


2.2 


3.1 


5.7 


8.1 


9.2 


13.0 


12.7 


17.9 


16.2 


22.9 


19.7 


27.9 


2.3 


3.2 


5.8 


8.2 


9.3 


13.2 


12.8 


18.1 


16.3 


23.0 


19.8 


28.0 


2.4 


3.4 


5.9 


8.3 


9.4 


13.3 


12.9 


18.2 


16.4 


23.2 


19.9 


28.1 


2.5 


3.5 


6.0 


8.5 


9.5 


13.4 


13.0 


18.4 


16.5 


23.3 


20.0 


28.3 


2.6 


3.7 


6.1 


8.6 


9.6 


13.5 


13.1 


18.5 


16.6 


23.5 


20.1 


28.4 


2.7 


3.8 


6.2 


8.8 


9.7 


13.7 


13.2 


18.7 


16.7 


23.6 


20.2 


28.5 


2.8 


4.0 


6.3 


8.9 


9.8 


13.9 


13.3 


18.8 


16.8 


23.7 


20.3 


28.6 


2.9 


4.1 


6.4 


9.1 


9.9 


14.0 


13.4 


18.9 


16.9 


23.9 


20.4 


28.8 


3.0 


4.2 


6.5 


9.2 


10.0 


14.1 


13.5 


19.1 


17.0 


24.0 


20.5 


29.0 


3.1 


4.4 


6.6 


9.3 


10.1 


14.3 


13.6 


19.2 


17.1 


24.1 


20.6 


29.1 


3.2 


4.5 


6.7 


9.5 


10.2 


14.4 


13.7 


19.4 


17.2 


24.3 


20.7 


29.3 


3.3 


4.7 


6.8 


9.6 


10.3 


14.6 


13.8 


19.6 


17.3 


24.5 


20.8 


29.5 


3.4 


4.8 


6.9 


9.8 


10.4 


14.7 


13.9 


19.8 


17.4 


24.6 


20.9 


29.6 


3.5 


4.9 


7.0 


9.9 


10.5 


14.8 


14.0 


19.9 


17.5 


24.7 


21.0 


29.7 


3.6 


5.1 


7.1 


10.0 


10.6 


15.0 


14.1 


20.0 


17.6 


24.9 


21.1 


29.9 


3.7 


5.2 


7.2 


10.2 


10.7 


15.1 


14.2 


20.1 


17.7 


25.0 


21.2 


30.0 


3.8 


5.4 


7.3 


10.3 


10.8 


15.3 


14.3 


20.2 


17.8 


25.2 


21.3 


30.1 


3.9 


5.5 


7.4 


10.5 


10.9 


15.4 


14.4 


20.4 


17.9 


25.3 


21.4 


30.2 


4.0 


5.7 


7.5 


10.6 


11.0 


15.6 


14.5 


20.5 


18.0 


25.5 


21.5 


30.4 


4.1 


5.8 


7.6 


10.8 


11.1 


15.7 


14.6 


20.6 


18.1 


25.6 


21.6 


30.5 


4.2 


5.9 


7.7 


10.9 


11.2 


15.8 


14.7 


20.7 


18.2 


25.7 


21.7 


30.6 


4.3 


6.1 


7.8 


11.0 


11.3 


16.0 


14.8 


20.9 


18.3 


25.9 


21.8 


30.8 


4.4 


6.2 


7.9 


11.2 


11.4 


16.1 


14.9 


21.0 


18.4 


26.0 


21.9 


31.0 



16 



GRAPHIC PRESENTATION SIMPLIFIED 



CHART 5. Too MUCH 
RULING. 



THOUSAND THOUS 


AND 
00 

80 
60 
40 
20 


































* 




. 


60 
40 
20 















; 










^^ 


^/ 


f 


J 




\j 










/\ 




/I 


V 


r\ 


[ 



















y 


X 




^ 


7 








j 























































J F M A M J JASONDJ F M A M J 
1946 1947 



CHART 6. RULING RE- 
PLACED BY TICKS. 



THOUSAND 
1100 



80 



60 



40 



20 



| | | | | | | I I I I I I I I I I In 
JFMAMJJASONDJFMAMJ 
1946 1947 



THOUSAND 
100 



80 



40 



20 




CHART?. COMBINA- 
TION OF TICKS AND 
RULING. 



I 1 J I I 



J F M A MJ JASONDJ F M A M J 




20 



CURVE CHARTS 



17 



ters up the chart unnecessarily and obstructs rather than facilitates 
reading the curve. 

Chart 6 shows the same curve on a grid in which the vertical and 
horizontal rulings are replaced by ticks. Those who favor this method 
hold that all ruling extending through the field of the grid detracts 
from the graphic effect of the curve. Against this view is the objection 
that the curves appear suspended in the air, separated from the scales 
so far that only a rough approximation of the amount value of any given 
plotting point can be made. The same difficulty is encountered in iden- 
tifying a point in the curve with the corresponding time unit below it. 

A compromise which has the sanction of generally accepted practice 
is shown in Chart 7. The ruling for the time scale is inserted at quar- 
terly periods, with the intervening months indicated by ticks. The 
amount scale is ruled at intervals of twenty, with ticks representing 
units of five. The grid field is fairly clear, and the identification of any 
point in the curve with the corresponding amount or date presents 
little difficulty. 

CURVE PATTERNS 

Chart 8 shows some of the types of curve patterns that are in com- 
mon use. The simpler designs are more economical from the point of 
view of the time consumed in the drafting and are generally the most 
effective, particularly in charts where the curves are close together. If 
the number of curves does not exceed three, marked contrast in the 





CHART 8. CURVE PATTERNS. 



18 GRAPHIC PRESENTATION SIMPLIFIED 

patterns selected is unnecessary. The usual practice is to employ a solid 
curve for the most important item, a dash curve for the next in order 
of importance, and dots for the third. Although somewhat less striking, 
different weights of solid curves may be used. On large display charts 
the curves may be fashioned of narrow strips of color. The strips are cut 
from the adhesive color sheets and pasted on the illustration board to 
form the line of the curve. Irregular curves should be lighter than 
smooth ones. 

Usually one of the curves in multicurve charts is of special signifi- 
cance. This should be plotted in a pattern that will stand out clearly 
from the others. For this purpose the line-and-dot and the ball-and- 
line types are very effective. 

OMISSION OF ZERO BASE 

An important rule in the drafting of curve charts is that the amount 
scale should begin at zero. In comparisons of size the omission of the 
zero base, unless clearly indicated, is likely to give a misleading impres- 
sion of the relative values and trend. The first chart on the opposite 
page shows the data plotted on a grid covering the full amount scale. 
In the second the amount scale under the lowest point of the curve is 
omitted. The third represents the generally approved method for indi- 
cating the omission of the zero base. Another form commonly used con- 
sists merely of a wavy line similar to the one at the bottom of the 
plotting field in the third chart. A light straight line with wavy sections at 
each end is sometimes used for this purpose. Whatever the device, it 
should clearly remind the observer that the full scale is not shown. 

Whether to include the full scale or omit part of it depends on the 
nature of the data. Fluctuations in the curve may occur within a lim- 
ited range only. Changes in prices of bonds are an example. Against 
the par value of $100, the price variation within a normal year may 
amount to only a few dollars. These small changes are important to 
investors, but if the whole scale from zero to 100 is included, they will 
be scarcely visible in the curve. 

100 PERCENT BASE IN INDEX-NUMBER CHARTS 

Exceptions to the zero-base rule are admissible in index-number 
curve charts, where the relation of the curves is to 100 rather than zero. 
The usual practice is to omit the zero base in such charts without the 



CURVE CHARTS 



19 



600 



500 



400 



200 



100 



HEAVY TANKS ACCEPTED 




1941 



600 



500 



400 



300 



200 



HEAVY TANKS ACCEPTED 






















/ 


/ 


1941 
















/r 


^. 


^ 


' 




\ 










/ 














) 


V 


y 


' 


















JFMA. MJJASOND 



JFMAMJJA50NO 



600 



500 



400 



300 



200 



HEAVY TANKS ACCEPTED 




JFMAMJJASONO 

CHART 9. OMISSION OF ZERO BASE. FROM Manual of Statistics, 
WAR DEPARTMENT, SOS. 



20 



GRAPHIC PRESENTATION SIMPLIFIED 



reminder of a break in the scale. Instead, the 100 percent line is made 
heavier than the other horizontal lines, as illustrated in Chart 10, to 
indicate that the curves are to be interpreted with relation to this base. 
A further reason for omitting the lower part of the amount scale in 
this chart was to gain space for the curve plotting. Had the full scale 



WAGES, PRICES, AND PRODUCTION 

MANUFACTURING INDUSTRY, 1935. 1936, 1937 



INDEX NUMBERS, MONTHLY AVERAGE 1935 '100 




National Industrial Conference Board 



CHART 10. 



CURVE CHARTS 



21 



been included, the curves would have been brought too close together 
for easy reading and comparison. The number of curves that can be 
shown effectively in a single chart, if they are very irregular or cross 
each other, is generally limited to three or at most four. But if the scale 
outside the plotting field is omitted, thereby increasing the separation 
between the curves, a fairly clear presentation of a larger number is 
possible. 

BREAK IN TIME SCALE 

Sometimes monthly time series start from a date several years subse- 
quent to the period with which they are compared. In such cases lack 
of space may make it advisable to omit part of the intervening time 
units. In Chart 11 index numbers of retail prices for fruits and vegeta- 
bles and for all foods during the last two war years are compared with 
the monthly average in 1939. To have included all the monthly data 



200 



180 



J40 



AVERAGE JFMAMJJASONDJFMAMJJASOND 




120 



100 



80 



CHART 11. RETAIL COST OF FOOD: 1939, 1944, AND 1945. 



22 GRAPHIC PRESENTATION SIMPLIFIED 

for the intervening four years would have made the chart too 
crowded. The monthly average in 1939 is connected with the plotting 
for the first month in 1944 by dash lines crossing the break in the time 
scale. 

If there is a gap in a series for the reason that the data are not avail- 
able, a break in the curve is necessary. The usual practice is to fill in the 
break with a light dotted line labeled "Data not available," or it may be 
footnoted, in which case the footnote symbol should be placed closely 
above or below the dotted line. 

FITTING CURVES TO GAIN SPACE 

When it is necessary to present a large number of series on one chart 
the curves may be "fitted" by the method illustrated in Chart 12. This 
method is feasible where the curves represent index-number series in 
which the base for all of them is the same. This is indicated in the 
example by the figure 100 at the first plotting point of each series. The 
spaces between the horizontal rulings are equal to ten points of the 
index. 

This chart was reduced from an original printed on a standard 8/2-by- 
1 1-inch sheet. Twelve series were fitted within this space. If drafted 
separately, each with its individual amount scale, they would have 
required many times the space, without the advantage of close com- 
parison secured by this method. 

Another feature of interest in this chart is the combination of 
yearly and monthly plotting. For the first six years the ball-and-line 
pattern is used for the yearly averages. This is continued through 
the following years, but with the addition of monthly data plotted 
in light dash curves. In this manner the monthly record is kept up 
to date until the end of the year, when the twelve-month average is 
added. 

Lettering the last figure at the end of the curve aids in showing 
the exact current status, which often is the point of chief interest 
to the observer. It is of special advantage to those who customarily 
think in terms of figures or who do not readily interpret charted statisti- 
cal data. In nearly all curve charts that picture extensive economic series 
it is advisable to label the high, the low, and the end points of the 
curve. 

Lettering of titles following the course of each curve is imposed in 
this chart by the close fitting. If there is sufficient space, titles should 



CURVE CHARTS 

DEPRESSION, RECOVERY, DEPRESSION 

UNITED STATES, 1929-1933 



23 



INDEX NUMBERS, I929-IOO 




93- 1935 1936 1937 1938 



National Industrial Conference Board 



CHART 12. 



be lettered horizontally. If title and curve are widely separated, a 
connecting arrow should be drawn between them. 

OMISSION OF TIME-SCALE RULING ABOVE CURVES 

In single-curve charts the time-scale ruling above the curve is often 
omitted. The employment of this method in charts containing a number 



24 



GRAPHIC PRESENTATION SIMPLIFIED 



of curves of approximately equal significance, as illustrated in Chart 13 
is sometimes criticized on the ground that the lack of ruling in the 
upper part of the chart tends to emphasize the importance of the high- 
est curve. 



Millions 
SO 




Type C 



V sf' 



Type D 



CHART 13. STOCKS OF C, D, AND K EMERGENCY RATIONS: 1943 (MILLIONS 
OF RATIONS). FROM Statistical Handbook, QUARTERMASTER CORPS. 

PLOTTING POINTS 

The data may be plotted on the vertical time-scale lines or centered 
in the spaces between them. Plotting to the line is generally preferred. 
The observer tends to read a curve where it crosses a line. Also, in 
plot-to-space charts ruled at wide intervals it is difficult to identify the 
plotting with the corresponding time unit at points where the angle of 
the curve does not change. 

Plot-to-space should be used if the series contains differing time 
units (such as end of the month and middle of the month) or if the 
data are reported at irregular intervals. 

Chart 14 shows both methods. The plot-to-space chart is ruled in 
quarterly subdivisions. The plot-to-line chart, having only eleven 
spaces, cannot be subdivided in this manner, as one of the quarters 
will be narrower than the other three. Plot-to-space must be used if 
quarterly or semiannual subdivisions are to be shown. 



CURVE CHARTS 



25 



100 



80 



60 



40 



20 



till 
PLOTTED -TO- SPACE 



A 





\ 



J F MAMJ JASOND 



100 



80 



GO 



40 



20 




I I I I 

PLOTTED- TO-LINE 



J F MAMJ JASOND 
CHART 14. PLOTTING TO SPACE OR LINE. 



26 



GRAPHIC PRESENTATION SIMPLIFIED 



MOVING-AVERAGE CURVES 

In Chart 15 the dash line represents a moving average of the monthly 
data. Moving-average curves are frequently added to indicate the 
approximate general trend, which is otherwise difficult to determine 
because of the irregular individual plottings. The solid curve in this 
chart shows the number of failures by months. The dash curve is the 
average, at successive entries, of the last twelve months. This average 
was recomputed each month by dropping the amount for the first 
month of the preceding twelve-month average and adding the data 
for the current month. 



1200 



1000 



800 



600 



400 



200 



BUSINESS FAILURES 




1941 1942 1943 1944 1945 1946 1947 

Cleveland Trust Company Business Bulletin 

CHART 15. 



CURVE CHARTS 



27 



Chart 16 shows another type of moving average, sometimes called 
a "progressive average," in which the additional curve represents an 
average of all the values to the date of each plotting. The amount for 
each year is averaged with all of the preceding entries. The lower curve 
represents an average of the placements to the current date. 

MILLIONS MILLIONS 

15 



12 



Yearly Placements 




I I 



1 I 



15 



12 



1937 1938 1939 I94O 1941 1942 1943 1944 1945 

CHART 16. PLACEMENTS BY U.S. EMPLOYMENT SERVICE: 1937 TO 1945. 



WEIGHT OF GRID LINES 

The grid lines should be lighter than the curves, with the base line 
somewhat heavier than the others. All vertical lines should be of equal 
weight, unless the time scale is subdivided in quarters or other time 
periods, indicated by heavier rules. Very wide base lines, sometimes 
employed for pictorial effect, distort the graphic impression by making 
the base line the most prominent feature of the chart. 



28 



GRAPHIC PRESENTATION SIMPLIFIED 



BROKEN CURVE 

If one curve in a chart of two or more contains a fluctuation so 
extreme that if plotted to scale it would dwarf the changes in the 
others, the curve may be broken as shown in Chart 17. When this is 
done, the point to which the curve would extend if shown in full should 
be lettered at the place where the break is made. 




10 



CHART 17. BROKEN CURVE. 
STOCK-PRICE CHARTS 

A curve type employed for recording the movement of stock prices 
is shown in Chart 18. The range of prices is indicated by heavy vertical 
lines at the various plotting points, extending from the highest to the 
lowest price at which the stock was sold. The intersecting curve shows 
the closing price for the day. 

Many stock charts include a column chart showing the volume of 
sales, placed below the price chart. 

CUMULATIVE-CURVE CHARTS 

The cumulative-curve type, which is illustrated in Chart 19, is used 
when the purpose is to show over-all results through an extended period 
of time. Each plotting represents the sum of all the entries up to date. 



CURVE CHARTS 



29 



40 
38 
36 
34 
32 
30 
28 
26 
24 
22 
20 



CLOSING 
PRICE 



DAYS 
MONTH 



DAYS 
MONTH 



CHART 18. STOCK-PRICE CHART. 



TITLE 




X 



LABEL 



-.--*' 



LABEL B 



TIME 

CHART 19. CUMULATIVE-CURVE CHART. 



30 



MULTIPLE-SCALE CHARTS 



GRAPHIC PRESENTATION SIMPLIFIED 



The use of two or more amount scales for comparisons of series in 
which the units are unlike and, therefore, not comparable for example, 
total wages paid compared with number of wage-earners, physical vol- 
ume of production with total value of production, prices compared 
with total sales generally results in an ineffective and confusing pres- 
entation which is difficult to understand and to interpret. Comparisons 
of this nature can be much more clearly shown by reducing the com- 
ponents to a comparable basis as percentages or index numbers. 

MULTIPLE-TIME-SCALE CHARTS 

This type, which is illustrated in Chart 20, is employed frequently 
for comparisons of monthly data by years. The time scale, in months, 
is lettered below the grid. The curves represent the years included in 
the comparison. 



TITLE 




TIME 

CHART 20. MULTIPLE-TIME-SCALE CHART. 
STAIRCASE-CURVE CHARTS 

The staircase-curve designation is applied to a type in which hori- 
zontal and vertical lines are substituted for slope curves. Chart 21 is an 
example. The employment of this form of curve is generally limited to 



CURVE CHARTS 



31 



charts of a single series. With two or more series the vertical lines of 
the curves frequently overlap. 



Bii: 

20 

15 

10 
5 


.Ions 
















APPROPRIATIONS 


i 














| 
























1940 










19 41 




: 1 




_: 










1917 


r ; 








1918 




.. rr ....J " 


I 




, , 



JFMAMJJASONDJFMAMJJASOND 

CHART 21. WAR APPROPRIATIONS: 1917-18 AND 1940-41. FROM Manual 
of Statistics, WAR DEPARTMENT, SOS. 

KEYS 

Keys for the identification of the curves in multicurve charts are 
seldom used or needed. Usually there is room to label the curves indi- 
vidually, without recourse to the key method of identification. The 
character and meaning of each curve should be instantly apparent. 
The use of keys hinders unnecessarily rapid reading and interpreta- 
tion of the chart. 



SURFACE-CURVE CHARTS 




charts can be dressed up 
by filling in the space between curves with shading or color. If adhesive 
patterns are used, the extra work required is insignificant in proportion 
to the improvement in appearance and effectiveness. 

DIFFERENTIATION OF TRENDS BY SHADING 

In Chart 22 the curves are supplemented with two patterns of shad- 
ing. The dotted area extends from the high level reached in 1937 to the 
lowest point of the downward trend. The following recovery move- 
ment, outlined by the curves, is filled in with diagonal shading. The 
original from which this chart was copied is in color, red denoting 
depression and blue recovery. 

A feature worth noting in this design is the economy of space gained 
by placing the titles within the grids. Another is the lettering of index 
figures at the points of greatest interest, which in this case are the turn 
of the trend and extent of recovery to the last date of the series. 

32 



SURFACE. CURVE CHARTS 



33 



DEPRESSION AND RECOVERY 

UNITED STATES, 1937-1938 



DEX NUMBERS, HIGH MONTH OF 1937-100 



DEBITS TO 

INDIVIDUAL ACCOUNTS 
(OUTSIOf NEW YORK CITY) 

I , , I , 




National Industrial Conference Board 



CHART 22. 



STRATA CHARTS 



The combination of surface with curves is especially effective in 
multicurve charts of the type shown in Chart 23, sometimes called 



34 



GRAPHIC PRESENTATION SIMPLIFIED 



"strata" or "band" charts. The width of each band represents the size 
of the amounts measured from the total of the amount plotted below 
it. The total curve is made heavier than the curves of the subdivi- 
sions. 

THOUSANDS 
1200 



1000 



800 



600 



400 



2OO 



1936 1937 1938 1939 1940 1941 1942 1943 1944 1945 

CHART 23. PRODUCTION OF MOTOR TRUCKS AND TRUCK TRACTORS: 

1936 TO 1945. 




DISTORTION PRODUCED BY STEEP CURVES 

If a chart contains a number of series which vary widely in individ- 
ual magnitude, optical distortion may result from the necessarily sharp 
changes in the angle of the curves. The space between steeply rising 
or falling curves always appears narrower than the vertical distance 
between the plotting points. This effect is illustrated in Chart 24, in 
which the vertical separation of the curves is the same at each plotting 
point. If the chart includes a number of series, the distortion may be 
increased by the cumulative irregularity of the curves. 



SURFACE. CURVE CHARTS 



35 




CHART 24. DISTORTION PRODUCED BY STEEP CURVES RISING OR FALLING 

AT A UNIFORM RATE. 



36 



GRAPHIC PRESENTATION SIMPLIFIED 



USE OF SHADING TO SHOW NET RESULTS 

Chart 25 illustrates the use of shading to show net differences in 
trends. The data consist of two series of index numbers covering the 
period 1939 to 1945, based on 1939 as 100. The upper curve represents 
an index of average weekly earnings of wage-earners in manufacturing, 
the lower a consumers' price index of commodities and services bought 
by families of wage-earners and moderate-income workers in large 
cities. The net gain in real earnings, after deducting the rise in prices, 
is indicated by filling in the space between the upper and lower curves 
with shading. 



200 



180 



160 




1940 1941 1942 1943 1944 1945 

CHART 25. AVERAGE WEEKLY EARNINGS OF WAGE-EARNERS IN MANUFAC- 
TURING, AND CONSUMERS* PRICE INDEX: 1939 TO 1945. 

SHADING SIGNIFICANT TIME PERIODS 

Shading is sometimes employed to call attention to those periods in 
the time scale which are of special interest or importance. Two patterns 
of shading are used for this purpose in Chart 26, in which the curve 
represents the amount of money in circulation per capita in each year 



SURFACE. CURVE CHARTS 

DOLLARS 
200 

180 




80 



60 



40 



20 



40 



20 



1916 '17 '18 '19 '20 '21 '22 '23 '24 '25 '26 '27 '28 '29 '30 '31 '32 '33 '34 '35 '36 '37 '38 '39 '40 '41 '42 '43 '44 '45 



CHART 26. MONEY IN CIRCULATION, UNITED STATES, PER CAPITA: 

1916 TO 1945. 



38 



GRAPHIC PRESENTATION SIMPLIFIED 



from 1916 to 1945. The figures for the end years of the two wars and the 
high and low of the intervening period are lettered below the curve. 

CONTRASTING NET CHANGES 

Different shadings or colors can be employed to contrast net changes 
where trends cross. The solid curve in Chart 27 represents the number 
of immigrant aliens admitted in the United States from 1932 to 1945. 
The departures over the same period are shown by the dash curve. The 
excess of departures over admissions, or net departures, is indicated by 
dotted shading. In 1936 the departures and admissions were approxi- 
mately equal. The cross-hatched area shows the volume of net admis- 
sions from that year on. 

THOUSANDS 



100 ft 




20 



1932 1933 1934 1935 1936 1937 1938 1939 1940 1941 1942 1943 1944 1945 

CHART 27. IMMIGRANT ALIENS ADMITTED AND DEPARTED: 1932 TO 1945. 
MINIATURE SURFACE-CURVE CHARTS 

Designs of the type illustrated in Chart 28, which involve the com- 
bination in a single chart of a considerable number of time-series 



SURFACE. CURVE CHARTS 



39 




CHART 28. TONNAGE RECEIVED AND SHIPPED BY QMC DEPOTS: 1944 
(THOUSANDS OF TONS). FROM Statistical Yearbook, QUARTERMASTER 

CORPS. 



40 



GRAPHIC PRESENTATION SIMPLIFIED 



miniatures, present somewhat difficult problems in arrangement and 
lay-out. Of these, the most complicated are the space adjustments that 
are necessary to make all of the columns the same height and provide 
for equal spacing between the miniatures while at the same time ad- 
hering to a uniform plotting scale. The chart shows the movement of 
supplies in twenty-two quartermaster depots during the year 1944. The 
dotted and diagonal shadings represent receipts and shipments respec- 
tively. 

LONG-RANGE BUSINESS-ACTIVITY CHARTS 

The design shown in Chart 29, "American Business Activity," is 
frequently employed for picturing economic movements over an ex- 
tended period of time. The base is a computed normal, designated by 
the abbreviation "Nor." The curve shows the deviations, in percents, 
above or below this level. 

This type is one of the most effective in presentations of data related 
to some central point, standard, or level, such as over or below an 
average, above or under 100 percent, in excess of or less than an esti- 
mate or forecast, above or below a given year. 




1940 1942 1944 1946 

Cleveland Trust Company Business Bulletin 

CHART 29. 



SURFACE. CURVE CHARTS 41 

EMPLOYMENT AND UNEMPLOYMENT 

UNITED STATES, 1929-1939 



EMPLOYMENT BY GROUPS 




1929 '30 '31 '32 '33 '34 '35 '36 '37 '38 '39 
AVERAGE 
ONTHLY 

National Industrial Conference Board 



CHART 30. 
SHADED STAIRCASE-CURVE CHARTS 

Chart 30 is an example of the use of shading in a staircase-curve 
design. It presents a large volume of information, compressed by in- 
genious designing into a small amount of space. Nearly all the lettering 
is within the plotting field. Two time scales at the bottom of the 
page are made to serve for all ten charts. 



42 



GRAPHIC PRESENTATION SIMPLIFIED 



SHADING VERSUS SOLID BLACK 

Chart 31 pictures the fluctuation in aerial activity on the French 
front during World War I. The use of solid black makes it difficult to 
identify points in the curve with the corresponding units in the time 
scale. Light transparent shading is preferable to black for charts in 
which, as occurs in this example, the variation is extremely wide. 



8,000 



2,000 




1,000 



1918 



1919 



CHART 31. HOURS SPENT IN THE AIR BY AMERICAN SERVICE PLANES AT 
THE FRENCH FRONT DURING WORLD WAR I. FROM The War with Ger- 
many, BY LEONARD P. AYRES. 



SURFACE. CURVE CHARTS 



43 



ECONOMY OF SPACE IN DESIGN 



Two measures of increase in volume of production in the United 
States, total and per capita of population, are shown in Chart 32. Effec- 
tive presentation of a number of charts in limited space is favored by 
the elimination of all scales and ruling not absolutely necessary, with 
labels lettered as closely as possible to the curves. If the curves repre- 
sent marked growth or increase, a single right-hand amount scale is 
sufficient. 

VOLUME OF PRODUCTION 
UNITED STATES, 1859-1934 



INDEX NUMBERS. BASE. 1899.100 

-1 400 




1859 '69 '79 '89 18991909 '19 '29'M 1859 '69 '79 '89 18991909 '9 2934 



MANUFACTURING 



MINING 




AGRICULTURE 



GENERAL 
CMANUFACTURING, MINING , 

AND AGRICULTURE COMBINED) 



National Industrial Conference Board 

CHART 32. 



IV 



RATE-OF-CHANGE CHARTS 




THER terms employed to des- 
ignate rate-of -change charts are "ratio," "logarithmic," and "semi-loga- 
rithmic." "Ratio" and "rate-of-change" are most commonly used. The 
latter has at least the advantage that it is clearly descriptive of what 
this design purports to show. 

GRID RULING 

Arithmetic and rate-of-change scale rulings are compared in Chart 
33. The spaces between the scale rulings in the arithmetic scale are 
equal. The corresponding spaces in the rate-of-change scale decrease 
in width with each successive ruling. 

The second chart shows an example of a three-cycle rate-of-change 
grid. The amount scale of the first cycle is 10 to 100; the second, 100 to 
1,000; and the third, 1,000 to 10,000. If a fourth cycle were required, 
the scale would be 10,000 to 100,000, and if a fifth, 100,000 to 1,000,000. 
In multicycle charts the amount scale in each succeeding cycle is ten 

44 



RATE-OF- CHANGE CHARTS 



45 



ARITH- 
METIC 



100 



RATE OF 
CHANGE 



x X 

x 

X 
X 

<b" ' 

/ 









10 



THREE-CYCLE 
GRID SCALE 



10,000 F 



1,000 



100 



10 



CHART 33. ARITHMETIC AND RATE-OF-CHANGE SCALES. 



46 



GRAPHIC PRESENTATION SIMPLIFIED 



times the scale in the next cycle below it. The number of cycles re- 
quired depends on the size of the component figures. As the rate-of- 
change design has no zero base, the bottom line is not, as in arithmetic 
charts, made heavier than the other lines of the grid. 

OMISSION OF SCALE RULINGS 

If space for the chart is limited, any part of the scale above or below 
the curve may be omitted. In Chart 34 the series covers a range of 400 
to 748. The scale begins at 400 and ends at 800. To have included in 
the scale the full cycle from 100 to 1,000 would have made the chart 
over three times larger. 




400 



CHART 34. TOP AND BOTTOM OF CYCLE OMITTED. 



RATE-OF-CHANGE CHARTS 



47 



RATE-OF-CHANGE VERSUS INDEX NUMBERS 

In Chart 35 the rate-of-change method for picturing rates of increase 
is compared with index numbers. The series represent the populations 
of Puerto Rico and Hawaii in the census years 1930 and 1940, with 
estimates by the Census Bureau for the intervening period. Reduced to 
thousands, the figures for Puerto Rico were 1,552 in 1930 and 1,877 in 
1940, and for Hawaii 368 and 426 in the same years. The first chart 
shows the data plotted on a rate-of-change grid. The difference in the 
slope of the curves is so slight that it is not clear which represents 
the higher rate of growth. The wide separation of the curves increases 
the difficulty of the comparison. 

The second chart shows the two series converted to index numbers, 
the base, 100, representing th,? population in 1930. In this foSrTi the 
comparison is much clearer and more easily understood. It shows not 
only which had the higher rate in growth of population, but also facili- 
tates close reading of the difference in rates of increase. 



THOUSANDS 
2000 



1600 

1200 
1000 

800 
600 



400 



300 



-RATE-OF-CHANGE 
- 
Puerto Rico 



Hawaii 



125 



120 



115 



no 



105 



100 



INDE) 


< NUM 


3ERS 














/ 




Puert 


o Rico 


/ 


/ 






/ 




f 




/ 


f 


f Haw 


iii 


/ 


1 


\ 


1 


1 



1930 1932 1934 1936 1938 1940 1930 1932 1934 1936 1938 1940 

CHART 35. GROWTH OF POPULATION, PUERTO Rico AND HAWAII: 

1930 TO 1940. 



48 



GRAPHIC PRESENTATION SIMPLIFIED 



METHODS FOR LAYING OUT GRIDS 

A simple method for laying out the grids for small rate-of-change 
charts is illustrated in Chart 36. After determining the vertical dimen- 
sion of the grid, secure a sheet of printed logarithmic paper on which 
the cycles are somewhat larger than this dimension. Place the printed 
cycle diagonally across the allotted vertical space and note the points 
for the ruling. If the series extends into numbers requiring more than 
one cycle, the others will duplicate the ruling of the first one. 

Grids up to six inches in height can be laid out from the scale in 
Chart 37. A piece of paper the length of the vertical dimension is 
placed over the scale and moved to the left until the same dimension 
is reached. The lines of the scale indicate the plotting points for the 
grid. Grids of display size can be laid out by extending the lines in 
Chart 37 on a larger sheet of paper. 




CHART 36. METHOD FOR LAYING OUT RATE-OF-CHANGE GRIDS WITH 
LOGARITHMIC PAPER. 

EXPLANATORY NOTE 
To a great many persons the difference between arithmetic and loga- 



RATE-OF-CHANGE CHARTS 



49 




CHART 37. SCALE FOR LAYING our RATE-OF-CHANGE GRIDS. 



50 GRAPHIC PRESENTATION SIMPLIFIED 

rithmic scale rulings is puzzling and confusing, particularly where both 
types appear in the same chart. In order to prevent any possibility of 
misunderstanding it is advisable to include in rate-of-change charts a 
note, preferably lettered within the grid, that will inform the reader 
that the relation between plotting points represents ratios of change, 
increase or decrease, rather than differences in size or amounts. A form 
sometimes used for this purpose in charts of this type is: "The scale 
focuses attention on rate of increase," or "decline," "loss," "growth," or 
"change." 

UNITS VARYING IN SIZE AND KIND 

Multicurve rate-of-change charts do not require the reduction of 
units of a different nature to a comparable basis. Chart 38 presents an 
example in which rates of increase of such diverse measures as miles 
of railway and telegraph and telephone wire, number of motor vehicles 
produced, value of radio apparatus, and production of airplanes are 
compared in the same chart. The unit quantities are adapted to the 
scale, which extends only to 10,000, although the actual totals of the 
basic units, with the exception of airplanes, all exceed this number. 
The units for railways and motor-vehicle production are in thousands; 
for radio apparatus, $100,000 value of production; and for telephones, 
10,000 miles of wire. 

LIMITED USE OF RATE-OF-CHANGE DESIGNS 

Notwithstanding the many efforts that have been made to popularize 
this method of graphic presentation, rate-of-change designs are seldom 
employed in statistical charts relating to business, industrial, and re- 
lated activities. The position taken by the statistical authorities in the 
procurement agencies of the Army during the last war typifies the pre- 
vailing attitude concerning the use of this graphic type. Standards of 
Presentation, issued by the Statistics Branch, Control Division, Army 
Service Forces, for the guidance of its staff in the preparation of statis- 
tical reports, contains the following suggestive comment on the use of 
logarithmic charts: 

"There are two dangers in using a log-scale chart: ( 1 ) that it will be 
used improperly, and ( 2 ) that it will be misread even if properly used. 
Therefore, although extremely valuable under appropriate conditions, 
this type is recommended only when the picture cannot be shown 



RATE-OF-CHANGE CHARTS 



51 



INDUSTRIAL CONTRIBUTIONS 

OF SCIENCE AND INVENTION 

UNITED STATES, 1830-1930 



00,000 MILES Or WIRE) 



RATE-OF-CHANGE CHART 



THE RATIO SCALE USED HERE FOCUSES 
ATTENTION UPON RATE OF GROWTH. 

THE UNIT FOR EACH CURVE IS INDICATED 
IN PARENTHESES. 



RAILWAYS 

0,000 MILES OPERATED) 



MOTOR VEHICLE 

PRODUCT ON 

[1.000) 



160 I860 1870 I860 1890 1900 1910 192.0 1930 




National Industrial Conference Board 

CHART 38. 



52 



GRAPHIC PRESENTATION SIMPLIFIED 



ARITHMETIC SCALE 

AMOUNT 
100 



LOGARITHMIC SCALE 

AMOUNT 
100 



J J F 

INDEX LOG. SCALE 



30% RATE OF / 
INCREASE / 




CHART 39. ARITHMETIC, LOGARITHMIC, AND INDEX LOG CHARTS. FROM 
Standards of Presentation, ARMY SERVICE FORCES. 

satisfactorily by some other means, such as conventional percentage 
measurement." 

These recommendations were followed by a graphic comparison of 
three methods, arithmetic, logarithmic, and index logarithmic, which 
is reproduced in Chart 39. 

The real test of these different methods is the impression the observer 



RATE-OF-CHANGE CHARTS 53 

gets from them. Viewing Chart A alone, he would probably conclude 
that the rate of increase in Curve 1 is greater than in Curve 2. Chart B 
shows which series increased at the faster rate, but affords no means for 
determining quickly by inspection the difference in the rates of in- 
crease. The comparison in Chart C is much clearer, with the advantage 
that the rates for both curves can be easily read from the scale. The 
light dash percentage-of-increase lines aid in reading the rates, but 
their inclusion in charts of more than two series is inadvisable, as the 
multiplicity of lines is likely to be confusing. 



,v 



BAR CHARTS 




HE term "bar charts" as used 
here applies to horizontal-bar charts. Vertical-bar charts, commonly 
designated "column charts," are discussed in another chapter. 

The bar type is one of the most frequently employed in graphic pres- 
entation and has many advantages. It is readily understood, even by 
those unaccustomed to reading charts, or who are not chart-minded. 
When the problem is one of comparing a large number of items, it is 
the only form that can be used effectively. It also possesses the out- 
standing advantage that it is one of the simplest and easiest to make. 

WIDE RANGE OF VALUES 

Chart 40 illustrates a problem frequently met with in the drafting of 
bar charts. The amount represented by the upper bar, Minneapolis, is 
nearly twice the next in size and around thirty times larger than the 
last item of the series. In order to gain space for the smaller bars of the 
series, the largest bar is broken. 

54 



BAR CHARTS 



55 



MINNEAPOLIS 



DULUTH 



KANSAS CITY 



ST. LOUIS 



CHICAGO 



OMAHA 



WICHITA 



PEORIA 



INDIANAPOLIS 



ST. JOSEPH 



MILWAUKEE 



SIOUX CITY 








CHART 40. RECEIPTS OF WHEAT AT PRIMARY MARKETS: 1943 
(MILLIONS OF BUSHELS). 

LETTERING VERSUS SCALES 

There is some difference of opinion among statisticians concerning 
methods for indicating the values in bar and column charts. Against the 



56 



GRAPHIC PRESENTATION SIMPLIFIED 



practice of lettering the figures on the bars, it is argued that the letter- 
ing interferes with the graphic comparison, and where placed at the 
end of the bars, adds to their length, thereby distorting the impression 
of relative values. Those who object to lettering the figures in this man- 
ner prefer the use of a scale placed over the upper bar. This method is 
illustrated in Chart 41. 

The real test of any method is how it functions in practical use and 
how far it meets the special needs of the person or persons for whom 
the chart is made. Assume, for illustration, that a display chart contain- 
ing the data on criminal offenses drafted in Chart 41 is prepared for a 
hearing composed of Congressmen, lawyers, and enforcement officers. 
They are not likely to know much about graphics, and generally they 
think of statistics in terms of numbers. Also, the facts presented in the 
chart involve matters requiring legislative or other remedial action, 
for which the actual figures are likely to be essential. 

Assume further, as an example of the type of difficulty often encoun- 
tered in the use of scaled charts, that in the course of the hearing some- 
one questions the accuracy of the high rate for aggravated assault in 



50 



100 



150 



200 



250 



South 
Atlantic 

East South 
Central 

Pacific 

West South 
Central 

East North 
Central 




UNITED 
STATES 

Mountain 

Middle 
Atlantic 



CHART 41. CRIME RATES IN URBAN COMMUNITIES (PER 100,000 POPU- 
LATION), BY GEOGRAPHIC DIVISIONS: 1943 (DATA SCALED). 



BAR CHARTS 



57 



the South Atlantic division, which greatly exceeds the rates for this 
offense in all the other divisions, and demands the exact data. Measured 
from the scale, the segment extends from a little beyond 50 to about 
half-way between 200 and 250. An estimate by the ordinary observer 
is likely to vary 10 to 25 percent from the true figure. Furthermore, 
under the conditions generally prevailing in hearings and conferences, 
both time and patience for estimates of this kind are lacking. 

Chart 42 presents the same data as Chart 41, but with the amount 
corresponding to each segment clearly lettered on the bars. The seg- 
ment labels are placed above the first bar. One disadvantage of scaled 
charts is the difficulty in labeling the segments so they can be easily 
identified. Unless the labels are inserted below the scale, which inter- 
feres with the use of the scale for measuring the size of segments and 
bars, they generally have to be lettered at irregular intervals in the 
blank space at the right of the bars. 



ROBBERY 



South 

Atlantic 

East South 
Central 

Pacific 




New 



England 



276 



CHART 42. CRIME RATES IN URBAN COMMUNITIES (PER 100,000 POPU- 
LATION ) , BY GEOGRAPHIC DIVISIONS I 1943 ( DATA LETTERED ON BARS ) . 

There is nothing exaggerated or unusual about the foregoing example. 
The writer has witnessed many like it in hearings and conferences with 
congressional committees. The presentation of scaled charts, without 



58 



GRAPHIC PRESENTATION SIMPLIFIED 



all the data readily available in worksheets or tables, nearly always 
resulted in requests for exact information that could not be immedi- 
ately supplied. 

In business conferences the chart may bear on matters of policy to 
be considered later by agencies or firms represented in the meeting. 
Yet if scaled charts are used to present the data, all the individual repre- 
sentative can carry away with him is his general impression of the 
graphic picture, with such approximations as he may have been able 
to make from the scale. 

In lettering figures on bar charts two rules should be observed. If the 
lettering is within the bar, it should be small enough so that a strip of 
the shading or color shows above and below it. If at the end, the letter- 
ing should be less in height than the width of the bar and slightly sepa- 
rated from it. This will obviate the possibility of the observer comparing 
the combined length of bars and lettering, rather than the bars alone. 





52 



CHART 43. LETTERING WITHIN AND AT THE END OF BARS. 

An example in which the first of these rules was not followed is shown 
in Chart 44. The lettering is so large that it cuts each bar in half. Also, 
it is centered on the bars, so that the figure on the first bar, 222, falls 
directly below 120 on the scale. Bars without subdivisions should be 
lettered near the end. In segmented bars the lettering may be centered 
on or placed at the end of the segments. 

No scale was needed in Chart 44. The use of both lettered bars and 
a scale in the same chart is confusing. The ruling under each stub item 
is unnecessary. Except in charts with very wide stubs, connecting 
lines from stub titles to bars are not required. 

ORDER OF COMPONENTS AND RULING 

Bar charts are sometimes arranged in the alphabetical sequence of 
geographical or other units, as, for example, the States or counties of a 



BAR CHARTS 



59 



State, but in this order are usually ineffective and hard to interpret. In 
nearly all cases a descending or ascending order of values will afford a 
clearer presentation and save the reader the effort of deciphering the 
relationship between components. 

Ruling of bar charts other than that required for outlining the bars 
may be entirely omitted. Or it may be held down to a single rule con- 
necting the left end of the bars. The question of more or less ruling is a 




CHART 44. POST, CAMP, AND STATION STOCKS OF NON-PERISHABLE SUB- 
SISTENCE ITEMS. FROM STATISTICAL REPORTS TO THE QUARTERMASTER 

GENERAL. 



60 



GRAPHIC PRESENTATION SIMPLIFIED 



matter of taste and judgment, depending chiefly on whether it aids in 
understanding and interpreting the statistics presented in the chart. 
The same criterion applies to the use of borders, a frame extending 
around the four sides of the chart, including the title and footnotes or 
sources below the grid. In crowded charts the margin between the 
border lines and the over-all chart area reduces the space available for 
the drafting. 

LETTERING STUB ON BARS 

In Chart 45 the names of the components, ordinarily placed as stubs 



FEDERAL EXPENDITURES 
PER CAPITA 

1946-47 FISCAL YEAR 



INT. FINANCE! 45 





GENL.GOVT. 
II 






36 



Cleveland Trust Company Business Bulletin 

CHART 45. 



BAR CHARTS 



61 



at the left of the chart, are lettered on and at the end of the bars. Ex- 
treme condensation is sometimes necessary in charts for publication, 
where there is insufficient space for the conventional arrangement. In 
this chart lettering the component titles in stub form would have short- 
ened the space for the bars by about one third. 

JOINED-BAR CHARTS 

Bar charts are easier to read and are of better appearance if space is 
left between the bars, but if the chart includes a large number of items, 
spacing may have to be omitted. Generally the values can be lettered 
within the bars. If the bars are too narrow to permit this, the choice is 
between lettering at the end or in a column at the left of the base line. 

Chart 46, copied from a joined-bar chart of the forty-eight States, 
compares the two methods. In the upper chart the figures are lettered at 
the right of the bars, and in the lower, between the stub items and the 
base line. The position of the lettering is a matter of choice, although 
the first method is more generally used. 



NEW YORK 

PENNSYLVANIA 

ILLINOIS 

CALIFORNIA 
OHIO 



1 9.9 



| 13.5 



7.9 



17.0 
I 6.9 



NEW YORK 13.5 

PENNSYLVANIA 9.9 
ILLINOIS 7.9 

CALIFORNIA 7.0 

OHIO 6.9 



12 



CHART 46. POPULATION OF FIVE LARGEST STATES: 1940 (IN MILLIONS). 




















q 
u 
























































o 




u 












i 


^ 






















































10' 
H 




1 










1 


~i 
i*> 


s 



























































* 

Q 












1^ 












































': 












O 10 
-< esj. 


> 


'I 








T 




g 






















^ 






















: 












1 


5 


V 

1 










/ 


c 




















T- 


% 






71 
















.*; 


^ 
X 










1 ^ 
O 

u. 




x \r>' 

Ul 

o 

Z 




< 


^ 
'^ 




^ 









Y////////////////////////////A 




V 

x 

v 




Y////////////////////////////A 


^ 


^ 












V/////////////////////////////J 


"." 


*r 




v////////////////////////////// 


r: 


- 




;r 


,. 


v 




Y/////////////////////////////A 













B ] 








* 




X 




'^ 




x 




^ 






.'. 


^ 




\ 




, 




x 




, 








i 




s 




' 












































x 








s. 
































































v 








v 
















x 












* 








, 




s^ 




* 




x. 




x 








^ 




s 




a 




\ 


*.* 


^ 








^ 




X 


V, 


, 


S 






' 













s; 




^ 








/ 






.;.' 


S 




"O> 




., 




\ 


V*. 


^ 








, 




x 
x 


'v 


\ 








. 








' 




\ 
\ 
x 




, 




^ 




'x 








-' 




x 

v 

x 

X 


V 







\ 

x 

\ 




; 






.\ 


^^ 






s 


^ 








! 








, 




\ 




* 




x 




/. 






P * B 






v 








\ 










\* 






>v 




"' 




























V 












* 




x 


1 


*, 




\ 


,' t 


* 









r '' 




N 




' 





















\ 




V 




^ 




* 








,* 




<N 


V 







N 


- . 








*.* 


^ 




V 





^^ 








" 








, 




x 

\ 




< 




^ 




^ 






,.' 


\ 




x 
x 
x 
s 










\ 
\ 


5 


'x 








< 




| 


>: 


; 




- 

Z 

o 






































































O. ^ 
LU O 

too 


ZJ 












C 
C 


s 

l 
J 






^ 


K 
K 

H 






ft 

*. 

V- 


\ 

> 






u 


3 






r- 


i 






r*> 

*~ 


4 






* 

s 


> 
-j 






u 

X 


i 

t 


ofc 


aE 

^ 






































































Q. 




















































</ 

t: 


) 
> 


















O 




V) 



































r> 














C 




j 


















m 




Q 


U 






















V 


) 






.- 
















c 


) 


















o 




13 
























a 


) 






^ 









































































X 


> 














i. 

























C3 
























r 


L 






C 

a 

a 


j 






^ 


'' 






-c 


1 






a 


^ 










Q 





























" 


1 








4 


j 




^J 
















. 


) 














r 
























t/ 


3 






(/ 


5 






^ 








c 


) 






c 












O 













c 
c 


T) 






fl 
c 
e 


[) 
7) 
9 






^ 




J 






- 
a 


- 


1 




^ 








c 



J 






a 

+- 
u 


j 

> 
} 















r 








. 








t 


X 






L 








*k 


v 






<^ 








V 


) 






.- 
















( 










> 














a 


> 






Q 


J 


3 




k_ 
















v 


7 














O 








c 

r- 
C- 


> 
) 






c 

L 


3 
T 

Ll 






2 


) 
3 






4- 




> 

! l 


3" 




2 

(S 








'a 

u 


) 
> 






* 
a 


J 









c 

. 






































&2 





































BAR CHARTS 63 

GROUPED-BAR CHARTS 

This type, illustrated in Chart 47, is employed frequently for pictur- 
ing changes in two or more categories from one time period to another 
or for comparisons with an estimate, average, or forecast. In Chart 47 
the diagonal bar represents the base of the index, 100, the dotted bar 
the September stocks, and the cross-hatched bar the forecast to March 
of the following year. While effective for comparing different aspects of 
a single component, comparisons between the components are difficult 
to interpret. In this example there are too many bars in each group and 
the groups are too close together. 

The use of grouped-bar charts should be limited, as a rule, to groups 
of two bars, with shading or color for each bar in marked contrast. The 
separation between groups should be around three quarters of the com- 
bined width of the bars. Groups of three bars or over present the eye 
with too many different comparisons at one time. 

KEYS 

If the bars include a considerable number of small segments, it may 
be necessary to use a key. This is usually placed at the right of the bars 
in the body of the chart, if space is available, or it may be lettered above 
the grid. Both methods are shown in Chart 48. 



TITLE 
WHEAT liiiliiiiiiiiii CORN R888&*! OATS 

Wtfftfa BARLEY PPIH ALL OTHERS 



BARS 




CHART 48. Two METHODS FOR PLACEMENT OF KEYS. 



64 



GRAPHIC PRESENTATION SIMPLIFIED 



The use of keys should be avoided as much as possible. Descriptive 
labels nearly always can be lettered over the first bar. Long labels may 
be lettered in two or three lines. 

PAIRED-BAR CHARTS 
This type is frequently used for comparisons of derived and basic 

MILES OF LINE NET INCOME PER MILE OF LINE 

::i|:;:!:;i;iii|:;i!i!!n;i| JAPAN 

10,469 




i33,500;i 



4,669 



GERMANY 
1935 



SWEDEN 
1935 



$315 



$177 



liilllllilllllllll 


AUSTRALIA 
(STATE) 
1935 

(FEDERAL) 
1935 


8..4S I 



DEFICIT PER MILE OF LINE 



$208 



$1,072 



NORWAY 

2,265 pi (STATE) 
1936 






'.".IE 



$296 








National Industrial Conference Board 

CHART 49. NET INCOME OR DEFICIT OF GOVERNMENTALLY OWNED OR 
OPERATED RAILWAYS IN VARIOUS FOREIGN COUNTRIES: 1935 OR 1936. 



BAR CHARTS 



65 



data. The purpose of the comparison in Chart 49 is to show the degree 
of success attained by various countries in the administration of govern- 
mentally owned or operated railways. The bars of the positive category, 
net income per mile of line, are arranged in descending order, from the 
largest to the least. The order of the negative category, deficits per mile 
of line, is reversed, the series beginning with Australia, which showed 
the smallest deficit, and ending with France, the largest. The bars on 
the left side of the chart indicate the size of the various railway systems. 



COMPLETION OF TERMINATION INVENTORY WORKLOAD 
AUGUST - NOVEMBER 1945 



DEPOT 


WORKLOAD 
1 AUG- 
30 NOV 

($000) 


PERCENT COMPLETED 


Total 
Philadelphia 

Jersey city 
Jeffersonville 

Chicago 
Boston 
Others 


206,344 
60,786 
11,440 
59,143 
61,429 
7,602 
5,994 




^HO^ 


JL*+-Auguat- October 




A 


l^^i^^^^i^^CNXVW^^J 

^^57 4^ Mm ^. 


Jk ^ 


^^p^ 


A 


?~;^ 


s 


K;^ 


A 


^^^wr?^ 


A 



CHART 50. CONTRACTOR-OWNED TERMINATION INVENTORY. FROM STA- 
TISTICAL REPORTS TO THE QUARTERMASTER GENERAL. 

BAR-AND-SYMBOL CHARTS 

A combination of bar and symbol is sometimes used in charts of 
grouped data. In Chart 50 the bar represents the percent completed 
through August to November. The pyramidal symbol indicates the cor- 



66 



GRAPHIC PRESENTATION SIMPLIFIED 



responding percent through August to October. Various types of sym- 
bols are employed. One frequently seen consists of a short vertical line 
intersecting the bar or placed at the right in line with it. 

The symbol device is usually less effective than grouped bars. Only 
a rough approximation of the amount indicated by the symbol is pos- 
sible, while a bar employed for the same purpose permits clearer plot- 
ting and lettering of the exact amount. Another difficulty is met with 
where the amounts represented by the symbols are greater than those of 
the bars. In such cases the symbol may be separated so far from the bar 



PERCENT OF TOTAL 
WORKING 




EXTRACTION 
OF MINERALS 
AND FORESTRY 



MANUFACTURING 
AND MECHANICAL 
INDUSTRIES 



NOT SPECIFIED 

INDUSTRIES 
X AND SERVICES 






PERCENT COVERED 



PERCENT 

NOT 
COVERED 



4.2 



29.4 




15.4 



2. 








TRANSPORTATION AND Q , 
COMMUNICATION **' ' 

C 

^DOMESTIC AND 

PERSONAL SERVICE 



PROFESSIONAL 
SERVICE 



AGRICULTURE, 
FISHING AND 
PUBLIC SERVICE 



23.8 




National Industrial Conference Board 

CHART 51. PROPORTION OF WORKING POPULATION COVERED BY THE OLD 

AGE PROVISIONS OF THE SOCIAL SECURITY ACT, USING THE DISTRIBUTION 
OF OCCUPATIONS OF THE 1930 CENSUS. 



BAR CHARTS 



67 



that the relation between the two is not clearly apparent. If there are a 
number of such items in the chart, the effect is extremely confusing. 

AREA-BAR CHARTS 

This form of 100 percent bar chart shows both the percentage dis- 
tribution of the categories and the relative size of the components. In 
Chart 51 the figures in the bars represent the percent in each group 
which are covered and not covered by the provisions of the Social 
Security Act. The figures at the left indicate the proportion of the total 
working force in each group. 

Unless the size of the component items is shown graphically in per- 
centage-distribution charts, it is advisable to include in a column fol- 
lowing the stub the figures from which the percents are derived. 
Comparisons of percentage distributions unaccompanied by the basic 
data do not provide sufficient information for a correct interpretation 
of the relative figures. 

OPTICAL DISTORTION IN BARS 



Care must be exercised in the choice of shadings to avoid combina- 
tions which produce optical distortions. Chart 52 shows the "bending" 
effect which results when two diagonal patterns are used and the appar- 
ent increase in the width of one segment when solid black and a hori- 
zontal line shading are combined. 

If bar charts are to be printed, the use of solid black in segmented 
bars should be avoided. Shading comes out lighter in printing than in 





DIAGONAL SHADING MAY "BEND" 
THE COLUMNS 



SHADING MAY AFFECT 
APPARENT WIDTH 



CHART 52. OPTICAL DISTORTIONS. 



68 GRAPHIC PRESENTATION SIMPLIFIED 

the drawing. When solid black and shading are used in the same bar, 
the black segment stands out more prominently than the shading, there- 
by overemphasizing the category it represents. 

INCLUSION OR OMISSION OF 100 PERCENT BASE 
It is desirable in percentage charts to show a measure which repre- 



r EM 

A 
B 
C 
D 


PERCENT ^ IC 
Perc 


)O 
ent 




140 




125 


94 | 




76 





ITEM 
A 
B 
C 

D 



PERCENT 



IOO 

Percent 
I 



ITEM 



PERCENT 



A 
B 
C 
D 


II 


9 


i 


4 



CHART 53. INCLUSION OR OMISSION OF 100 PERCENT BASE. 



BAR CHARTS 



69 



sents the 100 percent base. If the figures approach or exceed 100, as in 
the upper example in Chart 53, the base may be indicated by a vertical 
dash line. If all the percentage figures are small, however, the 100 per- 
cent horizontal scale is likely to leave too little space for an effective 
comparison, and in such cases it should be omitted. For example, the 
inclusion of a base line in the second chart dwarfs the length of the bars. 
In the lower chart the base has been omitted, thus permitting the full 
width of the chart to be utilized for a clearer comparison of the per- 
centage figures. 

DEVIATION-BAR CHARTS 

This design is frequently used for comparisons of positive and 
negative quantities, increases and declines, losses and gains, or de- 
viations from a standard or norm. The bars extend to the left and 
right of a vertical base line. The "positive" bars are generally arranged 
in descending order of sequence and the "negative" bars in ascending 
order. 

Chart 54 illustrates a problem frequently encountered in charts of 
this type. The length of the negative "Subsistence" bar imposes a wide 
separation between the bars and stub titles of the other items of the 
series. This can be overcome by lettering the stub titles opposite the 



GROUP OF 
ITEMS 


PERCENTAGE CHANGE, JANUARY TO FEBRUARY 1944 


DECREASE 


INCREASE 


All QM Items 
Equipage 
General Supplies 
Textiles 
Clothing 
Subsistence 




ill 

HO.. 

ange 




No ch 


^3W>^^^-^^^^^^^^^ 

KV, '* ,Vi', ViVxViV^VxV*', '.V.V, ,*' V.VAVVAVx'.'.'. ' 





CHART 54. PRICE CHANGES. FROM STATISTICAL REPORTS TO THE QUARTER- 
MASTER GENERAL. 



70 



GRAPHIC PRESENTATION SIMPLIFIED 



bars, within the body of the chart, as shown in Chart 55. By this 
method the stub titles are placed closer to the bars and more space 
is available for the graphic comparison. 



ARMY FUNDS UNCOMMITTED MILLIONS 

FEB 28, 1942, AND JAN 31. 1942 



DECREASE 



INCREASE 



Air 
Chemical Warfare 

Signal 168 
Medical 1 19 



1,889 




164 ^| Quartermaster 
561 BfHHJI Engineers 
Ordnance 



CHART 55. FROM Manual of Statistics, WAR DEPARTMENT, SOS. 
BASE FIGURES IN CHARTS OF RELATIVE NUMBERS 

In bar charts of averages, percentages, or other relative numbers, the 
item representing the combined figure for the whole series may be set 
off, as' in Chart 56, by larger lettering. Various other methods are em- 
ployed to distinguish it from the accompanying items: wider space 
above and below its bar and stub, a different shading or color, a bar 



BAR CHARTS 



71 



CONSUMERS' PRICES 

% INCREASE, MAY 1947 OVER 1935-39 AVERAGE 



FOOD 



CLOTHING 



HOUSE 
FURNISHINGS 



ALL ITEMS 



MISCELLANEOUS 



FUEL, ICE, 
ELECTRICITY 



RENT 




88% 



84% 



82% 



56 /o 



39/o 



!8/o 



I 



Cleveland Trust Company Business Bulletin 

CHART 56. 

wider than the others; or by placing it at the beginning of the series in- 
stead of in order of size. A common practice, if the last method is used, 
is to extend a vertical dash line from the end of the "All items" or 
"Total" bar through the bars below it. 

ADJUSTMENT OF STUB TO SPACE 

One of the faults most frequently observed in bar charts results from 
the common practice of fitting the stub space to the longest component 



72 



GRAPHIC PRESENTATION SIMPLIFIED 



title. As the bars constitute the graphic part of the chart, as much as 
possible of the horizontal space should be allotted to them. Long titles 
should be shortened by editing and abbreviation and, if necessary, by 
lettering in two or more lines. 

Chart 57 shows a comparison of two charts with identical stubs, 
the first with all item titles lettered in single lines, and the second 
with the longer titles lettered in two lines. The effect of shortening 
the stub of the lower chart in this way increases the space for the 
bars nearly one half. The bars are brought closer to the item 
titles, with marked improvement in the general appearance of the 
chart. 



GROUP 



EMPLOYMENT 




Manufacturing 

Trade 

Government: Federal, State and Local 

Finance, Service and Miscellaneous 

Transportation 



Government : Federal , 
State and Local 



Finance, Service 
and Miscellaneous 




CHART 57. ESTIMATED NON-AGRICULTURAL EMPLOYMENT, FIVE LARGEST 
INDUSTRIAL GROUPS: LAST QUARTER, 1945 ( IN THOUSANDS). 



BAR CHARTS 



73 



STEP-BAR CHARTS 

Chart 58 shows a design sometimes used instead of the conventional 
stub and bar arrangement for picturing the subdivisions of a total if 
their number is too large for an effective presentation in a single bar. 
The relation of the subdivisions to the total is somewhat clearer than if 
the bars were placed immediately following the stub titles. 



EUROPE 



NORTH 
AMERICA 



ASIA 



SOUTH 
AMERICA 



AFRICA 



OCEANIA 




CHART 58. EXPORTS (INCLUDING REEXPORTS), UNITED STATES, BY CON- 
TINENTS: 1940 (IN MILLIONS OF DOLLARS). 



UJI- 

o:c/) ( 



(/) 



. 
000 




74 



BAR CHARTS 75 

SLIDING-BAR CHARTS 

A variation of the deviation-bar type, sometimes designated "sliding- 
bar," is shown in Chart 59. It presents a large volume of data condensed 
without sacrifice of simplicity and clarity within a small amount of 
space. Economy of space is secured partly through the elimination of 
all scales and ruling not absolutely essential and partly through lettering 
the stubs, in this case years, over the base line of the bars instead of at 
the side of the chart. 

The original from which this chart was copied was in two colors, blue 
in two shades for investments and red in three shades for loans. Color 
offers many advantages over black-on-white shading for charts of multi- 
segmented bars. 



COLUMN CHARTS 




'OLUMN charts are bar charts 
arranged vertically. Their most frequent use is for picturing com- 
parisons of similar components at different times, while bar charts are 
generally employed to compare different components at the same time. 
The column design is particularly effective for the presentation of series 
which comprise a small number of time periods with few subdivisions 
of value. It is not well suited for comparisons of several time series nor 
for those which cover an extended period of time and have many 
plottings. 

COLUMN AND SEGMENT LABELS 

Column charts are considerably more difficult to design than bar 
charts, chiefly because of the problems they present in labeling seg- 
mented columns. Segments in bar charts can be clearly labeled over the 
first bar, while in column charts the labels are lettered across the col- 
umns or at the left or right side of the chart. Lettering of segments on 

76 



COLUMN CHARTS 



77 



the left side is usually not practicable if the first column is much shorter 
than the others nor on the right side if the chart is to be brought up to 
date at recurring intervals. In the latter case the labels, lettered on 
separate strips of paper, may be pasted with rubber cement at the right 
of the last column and shifted for future entries. 

Chart 60 illustrates the method commonly employed for segment 
labeling at the left of the chart. The dash lines connecting the segments 
should be included in column charts, unless the spaces between the 
columns are very narrow or there is marked irregularity in the size of 
the segments. 



Total 
during month 

Disposal*, re- 
distributions 
and transfers 



On hand at_ 
end of onth 




\\...r 



v \ 22,733 



24,010 

m 



SEP 



OCT 



CHART 60. DISPOSITION OF SURPLUS PROPERTY: JUNE TO DECEMBER, 
1944 (THOUSANDS OF DOLLARS). FROM Statistical Yearbook, QUARTER- 
MASTER CORPS. 

COMPARISON OF COLUMNS WITH CURVES 

Since the introduction of adhesive shading and color, the column type 
has supplanted curves to a considerable extent in charts of time series. 
This has come about chiefly because making a column chart with the 
new material represents little if any more work than would be re- 
quired for drafting the same data as curves, while improved appear- 
ance and more striking contrast can be secured through use of color or 



78 



GRAPHIC PRESENTATION SIMPLIFIED 



MILLIONS 
OF DOLLARS 



Other 
Sales 



Automobile 
Sales 



Loans 




6000 



5000 



4000 



3000 



2000 



000 



1939 1940 1941 1942 1943 1944 1945 

CHART 61. INSTALMENT CREDIT: 1939 TO 1945 (COLUMN DESIGN). 

shading. To the non-statistical observer a series in columns appears 
simpler and easier to understand than the same data plotted in a curve 
design. 

Charts 61 and 62 illustrate some of the advantages and disadvantages 



COLUMN CHARTS 



79 



MILLIONS 
OF DOLLARS 

2400 



2000 



MILLIONS 
OF DOLLARS 

~*2400 



2000 



1600 



1200 



800 



400 




1600 



1200 



800 



400 



1939 1940 1941 1942 1943 1944 1945 
CHART 62. INSTALMENT CREDIT: 1939 TO 1945 (CURVE DESIGN). 

of each type. The columns compare more effectively than the curves the 
amounts of credit extended in each category but, because two of the 
segments have no common base, do not indicate as clearly the trend or 
direction of change. For the latter purpose, in most cases the curve 
chart, in which all the curves are plotted from the same base, is gen- 
erally preferred. 



80 



GRAPHIC PRESENTATION SIMPLIFIED 



TO France 




MAY JUM JUL AUG SOTOCT MOV DEC JAT1 FEB KAftAPR MW JUfl JUL AU6 5EPTOCT HOX DEC JAfl FEB MIR APR HAY JUM 
1917 1918 1919 

CHART 63. MEN SAILING EACH MONTH TO FRANCE AND HOME: 1917 TO 
1919. FROM The War with Germany, BY LEONARD P. AYRES. 

HOLLOW COLUMNS 

Hollow columns representing simple comparisons of size can be used 
effectively in charts intended for printed reproduction, as illustrated in 
Chart 63. The ruling and lettering of the black portion of the chart were 
done with white ink. The columns are so narrow and the figures so large 
it was necessary to disregard the rule against vertical lettering. Single 
letters might well have been substituted for the three-letter abbrevia- 
tions in the time scale, which is too crowded for easy reading. 

COMBINATION OF SPACED AND JOINED COLUMNS 

The spaced columns in Chart 64 represent yearly averages from 1939 
to 1944. The joined columns show the data by months in 1945. The 



COLUMN CHARTS 



81 



300 



300 



250 



200 



150 




100 



1939 1940 1941 1942 1943 1944 J 



F M A M J J A 
1945 



S 



CHART 64. INDUSTRIAL PRODUCTION, MANUFACTURES: 1939 TO 1945 
(1935-39 AVERAGE = 100). 

average is indicated by the dash line intersecting the monthly columns. 
Amount-scale ruling was necessary, as the monthly columns are too 
narrow for lettering the values. For the same reason the months in the 



82 



GRAPHIC PRESENTATION SIMPLIFIED 



time scale are represented with single letters instead of the three-letter 
abbreviations usually employed for this purpose. 

The choice between spaced and joined columns depends on the 
length of the time series. In long time series spacing necessarily is 
omitted. In short series fairly wide spacing between columns, at least the 
width of the columns, gives a better appearance than narrow spaces and 
wide columns. 

COMBINATION OF CHART AND TABLE 

If the columns are too narrow for lettering, as in Chart 65, a table 
containing the data may be lettered within the grid field. This method 
is limited to small charts for publication or for desk use. It is not usually 



INDUSTRIAL PRODUCTION 





1937 1938 1939 1940 1941 1942 1943 1944 1945 



Cleveland Trust Company Business Bulletin 

CHART 65. 



COLUMN CHARTS 



83 



practicable in display charts. If the lettering is large enough to be 
legible at some distance, the table will take up too much space. 

GROUPED-COLUMN CHARTS 

This type is fairly effective if the number of columns in each group 
does not exceed two. Groups of three or more complicate the compari- 
sons between the categories in the various groups to the extent that 
interpretation of the chart becomes difficult. The columns should be 
clearly differentiated by contrasting colors or shading. 

In Chart 66 keys are employed to identify the categories represented 
by the columns. The use of keys should be avoided, but the alternative 
in grouped-column charts labeling the columns in the space above the 
groups is equally objectionable. Key labels may be omitted, as in this 



600 



500 



400 



300 



200 



100 



HEAVY AND LIGHT BOMBERS ACCEPTED - 1942 




JAN 



FEB 



MAR 



APR 



MAY 



JUN 



CHART 66. FROM Manual of Statistics, WAR DEPARTMENT, SOS. 



84 



GRAPHIC PRESENTATION SIMPLIFIED 



example, by placing the key directly under the corresponding designa- 
tion in the title. Another method, if color is used, is to letter the name of 
each component or category in the title in the same color employed to 
designate it in the chart. 

In Chart 67 one category is represented by columns and the other by 
circles connected with lines. This method permits clear lettering of each 
category. It is most effective for comparisons in which the values do 
not vary widely in size. 



Par. 




1.8 




2.3! 



2.5 



3.6 




2.9 



:2.5: 



1938 1939 1940 1941 1942 1943 1944 1945 



CHART 67. PAR AND MARKET VALUES OF BONDS SOLD ON NEW YORK 
STOCK EXCHANGE: 1938 TO 1945 (BILLIONS OF DOLLARS). 

COMBINATION OF COLUMNS AND CURVES 

An example of joined columns combined with a curve and surface 
overlay is -shown in Chart 68. The solid curve represents expendi- 
tures, and the ball-and-line curve receipts. The deficits from 1917 to 
1920, when expenditures exceeded receipts, are indicated by dotted 
shading. The space corresponding to the surplus from 1920 to 1930, 



COLUMN CHARTS 



85 



RECEIPTS, EXPENDITURES, AND NATIONAL DEBT 
FEDERAL GOVERNMENT, UNITED STATES, 1 91 3-1 936 

FISCAL YEARS ENDING JUNE 30 



EXPENDITURES 
RECEIPT^ 




I9I3'I4 15 16 17 18 *I9'20'2I '22'23'Z4. '25*26 '27 *28*Z9 30 '31 

National Industrial Conference Board 

CHART 68. 



86 



GRAPHIC PRESENTATION SIMPLIFIED 



when receipts exceeded expenditures, is filled in with a heavy black 
shading. Yearly deficits began again in 1931 and continued through 
1936. 

The amount scale is repeated at the right of the chart. This is usually 
desirable in charts covering long periods of time. The gross debt totals 
are lettered at the points of chief interest, the peak after World War I, 
the lowest point reached in the following eleven years, and the end year 
of the series. 

CHARTS SHOWING BOTH BASIC DATA 
AND PERCENTAGE DISTRIBUTION 

In charts where the purpose of the presentation requires that both the 
basic data and the percentage distribution appear, various methods are 
used to show both sets of figures. A small table of the percents or the 
amounts, whichever is least important, may be attached to or lettered 
on the chart. This procedure is seldom used, however, as it does not 
tie the series together closely enough for convenient reference and 
comparison. Another method commonly employed, illustrated in Chart 
69, consists in lettering both series on the columns. The basic figures are 
lettered above or within the columns and the percents below in smaller 
lettering, followed by the percent sign. The figures may be further dif- 
ferentiated by enclosing the percents within parentheses. Or the per- 
cents may be lettered immediately above the base line or centered on 
the columns. 



24O 




24O 

(18%) 




24O 




240 


18% 


18% 






"~ L 










18% 



X 

\> *V CHART 69. METHODS FOR LETTERING BASIC AND PERCENTAGE FIGURES ON 

N^^ COLUMN CHARTS. 

X 



COLUMN CHARTS 



87 



The design shown in Chart 70 is frequently employed where a 
graphic comparison of the basic data with the percentage distribution 
is desired. The lettering of the segment labels between the charts, in- 
stead of at the left of each of them, one set of labels thus serving for 
both charts, saves space and facilitates the comparison. 



AMOUNT 
(Millions) 



4.239 



Total 



PERCENT 

DISTRIBUTION 

100.0 100.0 




1932 
CHART 70. 



1945 



1932 



1945 



INTEREST PAYMENTS ON DEBT OF FEDERAL, STATE, AND LOCAL 
GOVERNMENTS: 1932 AND 1945. 



GRAPHIC PRESENTATION SIMPLIFIED 



DEVIATION-COLUMN CHARTS 

This type, consisting of columns above and below a common base 
line, is used, like the deviation-bar chart, for comparisons of positive 
and negative values and variations from a standard, estimate, or fore- 
cast. 

Charts 71 and 72 show two arrangements of the time scale. In the 
first, space sufficient for lettering the years forms the base separating 
the columns. In the second, the years are lettered below the chart and 
connected with the columns by dotted lines. 




Exports 







1915 



1925 



!935 



1945 



Imports 



1,674! 



i 2,047; 



4,227 



CHART 71. TOTAL EXPORTS AND IMPORTS OF MERCHANDISE: 1915, 1925, 
1935, AND 1945 (MILLIONS OF DOLLARS). 



COLUMN CHARTS 



89 



If the chart is for printed reproduction, the figures of the time scale 
may be pasted on the base line between each pair of columns. The let- 
tering should be small enough to leave a strip of shading or color be- 
tween the figures and the column rulings. If the difference in the size of 
columns is extreme, as in Chart 72, lettering on the base line is imprac- 
ticable. 



DOMESTIC CORPORATE SECURITY ISSUES 

UNITED STATES, 1919 - 1937 

COMPARISON OF NEW CAPITAL ISSUES WITH REFUNDING ISSUES 




1923:1924 1925 ! 1920 1927 1928 1929 1930 1931 ! 1932 I 1933 i 1934 I 1935 1 I93 i 1937 



National Industrial Conference Board 



CHART 72. 



CIRCLE CHARTS 




'OMPARISONS of size are much 
easier to interpret correctly from bars or columns than from circles, 
squares, cubes, or spheres. The use of various geometric forms for 
graphic purposes is compared in Chart 73. The second column in 
the first chart is exactly twice the height of the first column, but if 
the same amounts are drafted in the form of circles or squares, the 
difference between them appears to be considerably less. This effect 
is still more noticeable in the comparison of the same set of figures as 
cubes or spheres. 

COMMON ERRORS IN DRAFTING CIRCLES 

There is probably a higher percentage of errors in the drafting of 
circle charts than is found in any of the other graphic types, chiefly 
because of the complicated procedure for calculating the diameters of 
circles. The rule is: Divide the number represented by the circle by 
.7854 and extract the square root of the quotient. Unless the worksheet 

90 



CIRCLE CHARTS 



91 



USE OF DIFFERENT GEOMETRIC FORMS 
IN COMPARING SIZES 



1 




^1 








| 


1 


PLANES PLANES 


PLANES PLANES 


PLANES 


PLANES 


IN NOV. IN MAY 


IN NOV IN MAY 


IN. NOV 


IN MAY 


5,000 10,000 


5,000 10,000 


spoo 


10,000 


LINES, BARS OR COLUMNS 


AREAS 


VOLUMES 


Easy to compare 


Difficult to compare 


Yery difficult 


to compare 


oO 


oO 


Q 


Q 


PLANES PLANES 


PLANES PLANES 


PLANES 


PLANES 


IN NOV. IN MAY 


IN NOV. IN MAY 


IN NOV 


IN MAY 


5,000 IOPOO 


5,000 10,000 


5,000 


10,000 


DIAMETERS OF CIRCLES 


AREAS OF CIRCLES 


VOLUMES OF SPHERES 


PROPORTIONAL 


PROPORTIONAL 


PROPORTIONAL 


TO NUMBERS 


TO NUMBERS 


TO NUMBERS 


.All circle types are difficult to compare and should seldom be used. 



CHART 73. FROM Manual of Statistics, WAR DEPARTMENT, SOS. 

shows the correct diameters or radii, the draftsman, who has probably 
forgotten the rule if he ever knew it, is likely to fall back on the easier 
but erroneous method of making the diameters proportional to the size 
of the numbers. 

Chart 73 shows an example of the distortion in circle charts where the 
diameters instead of the areas of the circles represent the numbers com- 
pared. The first chart in the lower row, in which the diameters of the 
circles are proportional to the numbers, greatly exaggerates the differ- 
ence between the amounts compared. 

SHORT-CUT METHOD FOR CALCULATING THE SIZE OF CIRCLES 

The use of the following method, combined with the table of square 
roots shown on page 93, will save time and tedious computation in 
calculating the size of circles. 



92 GRAPHIC PRESENTATION SIMPLIFIED 



RULES 

A. Divide the smaller number by the larger. Express the quotient in a decimal 
of not more than two digits. 

B. Extract the square root of this quotient. (Table III lists the square roots of 
decimals from .01 to .99 to three places. ) 

C. Multiply the square root by the predetermined diameter of the larger circle. 
The product will be the diameter of the smaller circle. 

Example: Larger number, 500; smaller number, 100. Diameter of 
larger circle, 6 inches. 

A. 100 divided by 500 equals .20. 

B. Square root of .20 equals .447. 

C. .447 multiplied by 6 inches, diameter of larger circle, equals 2.7 inches, the 
diameter of the smaller circle. 

Large numbers should be rounded to not more than three digits. 
This will shorten and simplify the calculation considerably, and the 
quotients will be the same, for drafting purposes, as if the full figures 
were used in the division. The extraction of the square root, which 
generally is the stumblingblock for the draftsman, is covered by the 
table. 

Table 'II illustrates the procedure for calculating the diameters of a 
number of circles representing figures that vary widely in size. The 
assumed diameter of the largest circle is 9 inches. 

TABLE II 
CALCULATION OF CIRCLE DIAMETERS 

Amounts B, C, D Di- Square Square Root 
Component Amounts Rounded vided by A Root Times 9 Inches 

A 126,765 127 

B 88,441 88 .69 .831 7.5 inches 

C 37,552 38 .30 .549 4.9 

D 25,224 25 .20 .447 4.0 

If a slide rule is used to compute the component ratios and multi- 
plication of the square roots, the calculation of the circle diameters 
for a chart of five to ten circles should not take more than fifteen min- 
utes. The diameters of the smaller circles (square roots multiplied by 
the diameter of the largest circle) should be rounded to one decimal 
place. Drafting to hundredths of an inch is impracticable. 



CIRCLE CHARTS 93 

TABLE III 
SQUARE ROOTS OF DECIMALS .01 TO .99 



Deci- 


Sq. 


Deci- 


Sq. 


Deci- 


Sq. 


Deci- 


Sq. 


mal 


Root 


mal 


Root 


mal 


Root 


mal 


Root 


.01 


.1 


.10 


.316 


.40 


.632 


.70 


.837 


.015 


.125 


.11 


.332 


.41 


.640 


.71 


.843 






.12 


.346 


.42 


.648 


.72 


.849 


.02 


.141 


.13 


.360 


.43 


.656 


.73 


.854 


.025 


.158 


.14 


.374 


.44 


.664 


.74 


.860 


.03 


.173 


.15 


.387 


.45 


.671 


.75 


.866 


.035 


.187 


.16 


.4 


.46 


.678 


.76 


.872 






.17 


.412 


.47 


.685 


.77 


.877 


.04 


.2 


.18 


.424 


.48 


.692 


.78 


.883 


.045 


.213 


.19 


.436 


.49 


.7 


.79 


.889 


.05 


.224 


.20 


.447 


.50 


.707 


.80 


.894 


.055 


.235 


.21 


.458 


.51 


.714 


.81 


.9 






.22 


.469 


.52 


.721 


.82 


.906 


.06 


.245 


.23 


.480 


.53 


.728 


.83 


.911 


.065 


.255 


.24 


.490 


.54 


.735 


.84 


.917 


.07 


.265 


.25 


.5 


.55 


.742 


.85 


.922 


.075 


.274 


.26 


.510 


.56 


.749 


.86 


.927 






.27 


.520 


.57 


.756 


.87 


.933 


.08 


.283 


.28 


.530 


.58 


.762 


.88 


.938 


.085 


.292 


.29 


.540 


.59 


.768 


.89 


.943 


.09 


.3 


.30 


.549 


.60 


.775 


.90 


.949 


.095 


.308 


.31 


.557 


.61 


.781 


.91 


.954 






.32 


.566 


.62 


.787 


.92 


.959 






.33 


.575 


.63 


.794 


.93 


.964 






.34 


.584 


.64 


.8 


.94 


.970 






.35 


.592 


.65 


.806 


.95 


.975 






.36 


.6 


.66 


.812 


.96 


.980 






.37 


.608 


.67 


.819 


.97 


.985 






.38 


.616 


.68 


.825 


.98 


.990 






.39 


.624 


.69 


.831 


.99 


.995 



94 



GRAPHIC PRESENTATION SIMPLIFIED 



NEW INDUSTRIES AND EMPLOYMENT 
UNITED STATESJ879 AND 1929 

AVERAGE NUMBER OF WAGE-EARNERS, MANUFACTURING INDUSTRIES, 

PER MILLION OF TOTAL POPULATION 

1879 

49,437 WAGE-EARNERS 




1929 

72,731 WAGE-EARNERS 



18 NEW INDUSTRIES 

9,243 

13% 



ALL OTHER INDUSTRIE 
3,4881 
87% 




INCREASE, 1929 OVER 1879 

23.294 WAGE -EARNERS 




National Industrial Conference Board 



CHART 74. NEW INDUSTRIES AND EMPLOYMENT (CIRCLE DESIGN). 



CIRCLE CHARTS 



95 



CIRCLES COMPARED WITH BARS 



A comparison of circles and bars for picturing differences in size, 
with sectors and segments for subdivisions within components, is shown 
in Charts 74 and 75. The upper circle in Chart 74 represents the num- 
ber of manufacturing wage-earners per million of population in 1879, 
the middle circle the corresponding number in 1929, and the lower cir- 
cle the increase that took place over the fifty years covered by the com- 
parison. The sectors show the proportions of the increase contributed 
by old and new industries. 

Chart 75 presents the same data in a bar design. It shows more clearly 
than the circle chart the comparison between the totals for the two 
years and the extent of the increase contributed by new and old in- 
dustries. Comparisons of size in the form of circles have the effect of 
lessening the apparent difference between the quantities compared. 
The amount represented by the 1879 circle, for example, appears to 
be not far below that shown by the 1929 circle, although the latter ex- 
ceeds it by over 47 percent. 



1879 



1929 



Increase 
1929 over 
1879 




49,437 



New In- 
dustries 



All Other Industries 



:9,243; 
13%: 



63,488 



87%: 



72,731 




:: 1 40 5 1 :::-. 



23,294 



CHART 75. NEW INDUSTRIES AND EMPLOYMENT (BAR DESIGN 



SECTOR CHARTS 




iMONG statisticians and drafts- 
men the common designation for a sector chart is "pie" chart. It is less 
effective than bar and column designs for accurate reading and inter- 
pretation, particularly if the series contains a considerable number of 
components or if the difference between components is slight. The com- 
ponents "Small Arms" and "Artillery" in Chart 76, which compares 
sector and bar designs, are an example. In the sector chart they appear 
to be almost the same size, while in the bar chart below the difference 
between them stands Out clearly. The eye can compare linear measure- 
ments more accurately than it can areas of sectors or arcs of a circle. 

ORDER OF SECTORS 

The standard practice in drafting sector charts is to begin with the 
largest sector at the central point of the upper half of the circle. The 
arrangement of sectors is from right to left in order of size. If the chart 
contains an "All Other" or "Miscellaneous" sector representing a num- 

96 



SECTOR CHARTS 



97 



DISTRIBUTION OF VALUE 
OF EQUIPMENT - BY MAJOR GROUP 



FIRE CONTROL 
EQUIPMENT 



ARTILLERY 



80% 



SMALL 
ARMS 




AUTOMOTIVE 
MATERIEL 



20% 



DISTRIBUTION OF VALUE 
OF EQUIPMENT - BY MAJOR GROUP 

PERCENT OFTOTAL 
20 40 60 



AUTOMOTIVE 
MATERIEL 



SMALL ARMS 



ARTILLERY 



FIRE CONTROL 
EQUIPMENT 



CHART 76. COMPARISON OF SECTOR AND BAR CHARTS. ADAPTED FROM 
Standards of Presentation, ARMY SERVICE FORCES. 



98 



GRAPHIC PRESENTATION SIMPLIFIED 



Semi- 

Manufactures 
23% 




Manufactured 
Foodstuffs 



Finished 

Manufacturers 

16% 



Semi- 
Manufactures 
23% 




CHART 77. IMPORTS, UNITED STATES, BY ECONOMIC CLASSES: 1942. 



SECTOR CHARTS 



99 



her of small components, this is placed last in the series of sectors, irre- 
spective of its rank in order of size. 

LETTERING OF SECTOR CHARTS 

Sector charts should not be lettered radially, following the angle of 
the individual sectors. The lettering of component titles should be hori- 
zontal, within the sector if not too crowded, outside the circle if the 
sector is too small to carry the lettering horizontally. The figures should 
be lettered, if possible, within the sectors. The two methods are illus- 
trated in Chart 77, which represents a percentage distribution, by eco- 
nomic classes, of United States imports in 1942. As the title of the last 
sector is too long for horizontal lettering within the circle, it is placed 
outside, with a connecting arrow to the sector. 

GROUPING OF SECTORS 

Grouping of sectors may be indicated in various ways. A curved line 
drawn outside the circle, extending between the limits of the group 
and lettered with the group designation, is commonly employed for 
this purpose. Another method, illustrated in Chart 78, consists of a nar- 
row band of shading placed around the grouped sectors. 



NUMBER 
TOTAL PROCUREMENT DEPOT AWARDS, 93.1 (Thousands) 



VALUE 
TOTAL PROCUREMENT DEPOT AWARDS, $1,437 (Millions) 




CHART 78. CONTRACT AWARDS TO SMALL FIRMS: 1944. FROM Statistical 
Yearbook, QUARTERMASTER CORPS. 



100 



GRAPHIC PRESENTATION SIMPLIFIED 



MINIATURE SECTOR CHARTS 

The sector type in miniature sizes may be used effectively for com- 
paring in a single chart a considerable number of components sub- 
divided into percentages or broken down on some other basis. The 
accompanying example, Chart 79, represents the troop distribution by 
source of sixteen National Guard divisions in World War I. The black 
portion of each circle shows the part of each division drawn from the 
National Guard, the shaded portion represents troops drawn from the 
National Army and other sources, and the unfilled gap the proportion 
each division was short of its authorized strength when it sailed from 
this country. The names of the States from which the National Guard 







32- 




33* 




34 th 




36* 





38* 




40" 




CHART 79. COMPOSITION OF NATIONAL GUARD DIVISIONS, WORLD WAR 
I. FROM The War with Germany, BY LEONARD P. AYRES. 



SECTOR CHARTS 



101 



contingents were drawn are lettered in white on the black sectors. 
The figures outside the circle to the left are the division numbers. 

SEPARATION OF SECTORS FOR EMPHASIS 

The importance of an individual sector may be emphasized by the 
method illustrated in Chart 80. This variation in design is little used. It 
is frequently criticized as tending to concentrate the attention solely on 
a single feature of the chart. 

CAUSES OF DELAYS IN MAY DELIVERIES 
SIGNAL CORPS - 1942 



OTHER 



SHORTAGES 
IN MATERIAL 



CONFLICTING 
CONTRACTS 





SHORTAGES 
IN EQUIPMENT 



PERSONNEL 
DIFFICULTIES 



CHART 80. ADAPTED FROM Standards of Presentation, ARMY SERVICE 

FORCES. 

OMISSION OF SHADING 



Sector charts do not require strong contrasts. Charts of this type with 
all shading omitted, although not as striking in appearance, are often 
quite as effective in conveying a correct impression of the relative values 
as would be secured with shading or color. If the titles are long, the 



102 



GRAPHIC PRESENTATION SIMPLIFIED 



lettering will occupy so much space within the sectors that most of the 
advantage of contrasting shading is lost. In Chart 81, for example, the 
titles of the fourth and fifth sectors are so long that nothing would be 
gained by surrounding them with narrow strips of shading. 

Overloading is the commonest fault in the planning and drafting of 
sector charts. If the number of sectors exceeds five or six, the lettering 
is likely to be crowded and confusing, with too many patterns of shad- 
ing or varieties of color. Simplicity, the first essential in all charts, is of 
special importance in designs of this type. 



PRO- 
FESSIONAL: 
PERSONS 

1.8 



UNSKILLED 
WORKERS 



SKILLED WORKMEN 
AND FOREMEN 
6.0 



PROPRIETORS, 

MANAGERS AND 

OFFICIALS 

8.7 



SEMI-SKILLED 

WORKERS 

7.3 




CHART 81. DISTRIBUTION OF MALE LABOR FORCE 14 YEARS OLD AND OVER, 
UNITED STATES: 1940 (IN MILLIONS). 

SECTOR TOTALS 

Totals of a number of components should not be represented by 
interior arcs drawn across the sectors. Chart 82, in which the lower arc 
represents the sum of the expenditures by the War and Navy Depart- 
ments, although the total corresponds in actual size to the larger arc of 
the circumference, is an example. Subdivisions of a total may be shown 



SECTOR CHARTS 



103 



by light dash lines, as in the second chart, or by different patterns of 
shading, but the total should be lettered outside the circle. 




TOTAL 

MILITARY 

SI, 368 




CHART 82. MILITARY EXPENDITURES, UNITED STATES: 1939 (MILLIONS 

OF DOLLARS). 



IX 



STATISTICAL MAP CHARTS 





lAP charts are employed to 
show the geographical distribution of statistical data. If they are to be 
small-sized, map charts are generally made with outline maps on which 
the boundary and division lines of the geographical units are printed. 
Outline maps on paper, ranging from letter size to several feet square, 
can be obtained from firms specializing in map production. The outlines 
for display maps on illustration board, where no copy of the size re- 
quired is available, must be drafted. This is a difficult job, but a photo- 
stat of a small map blown up to the dimensions of the display outline 
will serve as a useful guide in the lay-out and drafting of the display- 
size map. 

About all a small map chart can accomplish graphically is to indicate 
the geographical distribution of the data, with a rough measure of the 
gradation in size of the numerical components. Attempts to show fine 
distinctions and small differences are nearly always ineffective. They 
only complicate and render more difficult a clear understanding of the 
general purpose for which the chart is planned. 

104 



STATISTICAL MAP CHARTS 



105 



USE OF COLORS 

Solid colors and colored shadings are generally used on maps not in- 
tended for printed reproduction, although the effect of gradual transi- 
tion from high to low volume or size of the categories is not as clear as 
with black shadings. Red is often used to designate the highest category, 
blue and green for the next two, followed by shadings of these colors, 
with yellow for the lowest category. Care should be taken to avoid harsh 
color contrasts. Light tints are more effective and have a more pleasing 
appearance than dark colors. 

MAPS WITH SURFACE DESIGNATIONS OF CATEGORIES 

The method most generally used for charting statistical data in maps 
is illustrated in Charts 83 and 84. Chart 84 is a section of a United States 




CHART 83. PERCENT OF DRAFTED MEN PASSING PHYSICAL EXAMINATION, 
BY STATES, WORLD WAR I. TAKEN FROM The War with Germany, BY 

LEONARD P. AYRES. 



106 



GRAPHIC PRESENTATION SIMPLIFIED 




1,000 to 4,000 
500 to 1,000 
100 to 500 
Vnder 1 00 



CHART 84. NUMBER OF CONTRACTS AWARDED IN EACH STATE IN 1944. 
FROM Statistical Yearbook, QUARTERMASTER CORPS. 



STATISTICAL MAP CHARTS 107 

map representing the distribution by States of quartermaster depot 
contract awards in 1944. The grouping of the data in six categories is 
indicated by the key at the lower right of the chart. 

This type of map chart is the clearest and easiest to read and costs 
the least in labor and time for drafting. As the adhesive colors and 
shadings come in large sheets, an application for a single category may 
cover half or, with small maps, all of the surface. The areas which are 
not in that particular category can be cut out in a few minutes. If figures 
for each geographical unit are shown, the color or shading is afterward 
cut out around them so they will stand out clearly. Or the figures, let- 
tered on strips of white paper, can be pasted in each geographical unit. 

The outline for Chart 84 was smoothed to save the time required for 
duplicating the exact configuration of the bays and inlets of the Atlantic 
Coast and the windings of State river boundaries. 

The use of white space to represent one of the categories is illustrated 
in Chart 83, page 105. It is fully as effective as the addition of a fourth 
shading would have been. 

DOT MAPS 

In Chart 85 the distribution of values is shown by small dots of equal 
size plotted by localities within the States. The drafting of a chart of 
this type is laborious and time-consuming. Each dot represents an indi- 




CHART 85. RURAL REHABILITATION CASES RECEIVING ADVANCES OF CAPITAL 
OR GOODS: 1935. FROM Rural Youth on Relief, WPA, DIVISION OF SOCIAL 

RESEARCH, 1937. 



108 GRAPHIC PRESENTATION SIMPLIFIED 

vidual plotting. The number of plottings in this chart runs to over one 
thousand. 

The use of dots on map charts is open to the objection that in locali- 
ties of high concentration, where the dots overlap, no variation in the 
data can be shown. The area becomes a black cloud on which no addi- 
tional plottings will register. 

MAPS WITH CIRCLES 

Chart 86 is a section of a map published by the Bureau of Agricul- 
tural Economics representing the distribution of Florida grapefruit in 
the 1932-33 season. The figures accompanying the circles add greatly 
to the clarity of the presentation. The circles alone, due to the difficulty 
of comparing circular areas, do not convey a clear impression of the dif- 
ference in the amounts. For example, the circles in Illinois and Ohio 
appear to be nearly the same size, yet the figure for Ohio exceeds that 
for Illinois by 40 percent. 

Objections to the use of circles for extended series in map charts are 
the length of time required for the calculation of the diameters and the 
close plotting of minute differences in size. Even employing the short- 
cut method described in Chapter VII, the various steps that must be 
gone through to compute the diameters of circles for forty-eight States 
consume several hours of time. The plotting and inking of forty-eight 
circles, if accurately done, takes at least as long. When the ineffective- 
ness of circles for comparisons of size is added, there is not much to be 
said in favor of their use in map charts containing a considerable num- 
ber of geographical units. 

The difference in the size of the States presents a problem in draft- 
ing United States map charts if the data are to be shown with circles. 
In most economic distributions the numbers for the New England 
and contiguous States are likely to be large. A circle which in the 
States of the South or West may be a mere dot has to be made so 
large in the lower New England States that it blots out all other 
details. Frequently the circles for Rhode Island, Connecticut, and 
Delaware must be placed at the right of the chart, outside the State 
boundaries. 

MAPS WITH COLUMNS, BARS, OR CURVES 

On map charts comparisons of size in the form of columns are diffi- 



STATISTICAL MAP CHARTS 



109 




Figures indicate 
number of cars 



U.S. Department of Agriculture 

CHART 86. DISTRIBUTION OF FLORIDA GRAPEFRUIT IN THE 1932-33 SEASON. 

cult to interpret, due to the lack of a common base. Time series shown 
as curves, columns, or bars are open to the same objection. They com- 
plicate the chart by presenting too many types of comparisons at the 
same time. The primary purpose of a statistical map chart is to show 
two things : geographical location and differences in size. The inclusion 



110 



GRAPHIC PRESENTATION SIMPLIFIED 



of additional aspects of the data, difficult to read and compare, tends to 
obscure a clear impression of the main objective of the chart. 

PIN AND TACK MAPS 

Geographical distributions on wall maps may be shown by means of 
map tacks or pins of various forms or colors. Strings of beads or colored 
yarn may be employed for connecting lines or for curve effects. The 
facility this method offers for quickly making changes and adding new 
data makes it especially adapted for a continuing record of changing 
conditions. No skill or knowledge of drafting is needed for the prepara- 
tion and maintenance of such charts. Figure 2 illustrates some of the 
map tacks in current use. 





Series I 
500 I 

The Most 

Common Size 
used is this 
spherical head 
pin. 



oo 

Series 
600 




For Congested 
Maps or charts 
and holding ter- 
ritory or route 
cord in place 
use this pin. 



Series! 
5400 

Glass Spot 

Pin . The spot 
is fused in the 
head of the pin. 
It cannot come 
off. 



Of 

Series I 
5000 I 

Enamel Spot 

or "Ring Pin." 
Spot is enam- 
eled on the 
glass. This 
and the glass 
spot pin pho- 
tograph as a 
ring. 



Series 
7000 

Oil Well Pin. Shape 
resembles oil well der- 
rick. Can be supplied 
with two-colored head. 




Seriu550 



Series4500 



Map Ring. Colored 
celluloid ring for slip- 
ping over head of pin 
to indicate additional 
fact. 




Pins and Beads 



An An Un Dn 

Series Series! Series I Series! 
6300 6100 I 6400 | 6200| 



Educational Exhibition Company, Inc. 

FIGURE 2. MAP TACKS FOR LARGE MAP CHARTS. 



STATISTICAL MAP CHARTS III 



FLOW MAPS 

This designation covers an extensive variety of designs and purposes. 
Some types, such as traffic maps, show direction of movement with con- 
necting lines. Both direction and volume may be shown by thickening 
the portions of the line or curve representing heavy volume of traffic or 
by repeating the curve in parallel or converging lines. The latter method 
is frequently used for charting ocean routes of travel or commerce. Map 
charts in which certain localities, cities, or centers of distribution are 
connected with other points on the map by single straight or curved 
lines to show a particular relationship are also classed as flow maps. 
Many of these types are simple in design and in drafting. Those in 
which the thickness of the curve or line represents varying magnitudes 
of the data are difficult to plot accurately, as the variations on such a 
small scale are necessarily minute. 

KEYS 

Nearly all map charts require the addition of keys to identify the 
colors or shadings in the chart. If the distribution of the data by size 
is markedly irregular, the selection of categories may have to be made 
without much regard for uniform gradation in amounts. The number of 
keys should be kept down, if possible, to not more than five or six. 



X 



PICTORIAL CHARTS 




tHE value of pictorial charts for 
presenting statistical facts depends chiefly on how and for whom they 
are used. They are well suited for illustrations in newspapers and 
magazines and for any type of audience requiring novelty in form of 
presentation to attract attention and interest. The use of pictorial 
figures does not give the best results, as a rule, in charts prepared 
for business - or professional men, who generally prefer the simpler 
conventional types of design, without ornamental or pictorial trim- 
mings. 

Pictorial charts may be used effectively, if they are accurately 
drawn, in showing simple comparispns of size. However, many pub- 
lished pictorial comparisons representing the human figure are mis- 
leading, due to the fact that in the drafting only one dimension, that 
of height, is taken into account, while the impression received by 
the observer is that of volume or bulk. Drawings of this kind, based 
on a single dimension, always exaggerate the difference between the 
numbers compared. 

112 



PICTORIAL CHARTS 113 



BAR PICTORIAL CHARTS 

A bar pictorial design frequently used is made up of rows of small 
figures or symbols equal in size, each representing a stated number of 
units. The chief problem in planning a chart of this type is the selec- 
tion of figures or symbols sufficiently representative of the components 
to make the association obvious to the observer. Otherwise they are 
less effective than the ordinary rectangular-bar designs. Examples as 
far-fetched as a spool of thread to represent the textile industry or an 
ax for the lumber industry are occasionally seen. 

Chart 87, representing the employee distribution in a large tele- 
phone company, is an example of the bar pictorial type. The figures 
in each bar suggest the type of occupation designated in the stub. 

NUMBER OF EMPLOYEES 



MALE 

PLANT 

CRAFTSMEN 



FEMALE 

TRAFFIC 

OPERATING 



CLERICAL 



ALL OTHER 



tttttitiiitl 

iimiminmt 




1 1 II I II! 



. 

EACH SYMBOL REPRESENTS IOOO EMPLOYEES. 

American Telephone and Telegraph Company 

CHART 87. TYPICAL DISTRIBUTION OF EMPLOYEES IN A LARGE TELEPHONE 

COMPANY. 

A large variety of pictorial symbols printed in sheets may be ob- 
tained from firms engaged in the production of material for this type 
of graphic presentation. A few samples are shown in Chart 88. Charts 
representing industrial data are sometimes made of small drawings of 
machinery, materials, or products. If special designs are required, they 



114 



GRAPHIC PRESENTATION SIMPLIFIED 



should be drawn oversize and reduced photographically for duplica- 
tion by photostat or offset printing. 












r//////A 

The Pictograph Corporation 

CHART 88. PRINTED SYMBOLS FOR USE IN BAR PICTORIAL CHARTS. 

Both the advantages and the disadvantages of the pictorial method 
are illustrated in Chart 89. The chart was copied from the first quar- 
terly report of the War Assets Administration in 1946. It is unquestion- 
ably much more striking in appearance than a plain bar chart of the 
same data would be. The pictorial figures for "Plants and Plant Sites" 
and "Aircraft" probably aid the reader to some extent in visualizing 
these components. It is doubtful whether the others add anything to 
the presentation. An undershirt is hardly representative of the thou- 
sands of miscellaneous items listed as consumer goods, nor a motor of 
the huge variety of surplus capital and producer goods. 



a 
s-l 



3 



o o co 

5 I 



m > 

-< 





O 



8 

O CO 



O 

TJ 

m 
3J 



115 



XI 



MISCELLANEOUS TYPES 




i HE types of design described in 
the preceding chapters comprise those that are most commonly used 
for the graphic presentation of statistical facts. The examples dis- 
cussed in this chapter include a number" of variations and adaptations 
of the basic types, employed to bring out clearly special aspects of the 
subject-matter. 

RANKING CHARTS 

Charts 90 and 91 present two methods for comparing series in 
which the components are arranged by order of their rank in a com- 
mon attribute. In Chart 90 a number of commodity groups are ranked 
at two different periods on the basis of wholesale-price levels. In the 
first column the groups are ranked on the monthly average in 1939, 
and in the second, on the price level in December, 1945. The change 
between the two periods is indicated by the connecting dash lines. 
This method is effective with short series, but if the chart includes a 

116 



MISCELLANEOUS TYPES 



117 



1939 
RANK AVERAGE 



1945 
DECEMBER RANK 



2. 



4. 



5. 



6. 



8. 



Hides and leather 
products 


X 

N j 

V""'-' 
- \ /> 

\ / 

\ 1 

\ \/ 

^ a %. ' 

x / V 

v . A 


Farm products 


1. 






Metals and metal 
products 


Building materials 


2. 






3. 

4. 


Building materials 


Hides and leather 
products 






House furnishing 
goods 


Foods 






Chemicals and 
allied products 


. \/ A 
\ >x' 


Metals and metal 
products 


5. 
6. 
7. 
8. 
9. 
10. 


\ / /V 


Miscellaneous 


> t' x 

\ y 


House furnishing 
goods 




\ A 

Y/ \ 
\ / * \ s 
\ / / \ \ S 
A / \ ^ 
/ \t \ s \ 




Fuel and lighting 


Textile products 






Foods 


/ /v\ ^ 

/v\ \ 

K V V 


Chemicals and 
allied products 






Textile products 


x-r \ \ 
/ \ 


Miscellaneous 




N 




Farm products 


Fuel and lighting 



10. 



CHART 90. RANK OF COMMODITY GROUPS BASED ON WHOLESALE PRICES: 
1939 AVERAGE AND DECEMBER, 1945. 

large number of components, the criss-crossing lines become too con- 
fusing for easy reading. 

In Chart 91 the States are arranged in order of total ranking in ten 
educational features. The range in each feature for the whole country 



118 



GRAPHIC PRESENTATION SIMPLIFIED 



1 WASHINGTON 

2 MASSACHUSETTS 

3 NEWM3RK 

4 CALIFORNIA 

5 CONNECTICUT 
b OHIO 

7 NtW JERSEY 

8 ILLINOIS 

9 COLORADO 
(Q INDIANA 

I \ RHODE, I5LAND 
12 VERMONT 

15 WW HAMPSHIRE 

14 UTAH 

15 OREGON 

16 MONTANA 

17 MICHIGAN 

18 N DAKOTA 

19 IDAHO 

CQ MINNESOTA 
i\ IOWA 

12- MAINE. 

3 PENNSYLVANIA 

tA KANSAS 

5 NEBRASKA 
ZO 5. DAKOTA 
& NEVADA 

ea VMCONSIN 

9 WYOMING 

30 ARIZONA 

31 OKLAHOMA 

x MISSOURI 

33 W VIRGINIA 

34 FLORIDA 

35 DELAWARE 
56 MARYLAND 

37 TEIMNE55LE 
58 TEXAS 

39 LOUISIANA 

40 NEW MEXICO 

41 VIRGINIA 

42 KENTUCKY 

43 ARKANSAS 

44 GEORGIA 

45 MIS\S3\Ppl 

46 N CAROLINA 
4? v> CAROLINA 
46 ALABAMA 






CHART 91. RANK OF EACH OF THE UNITED STATES IN TEN EDUCATIONAL 
FEATURES: 1910. FROM GREENWICH EDUCATIONAL SURVEY. 



MISCELLANEOUS TYPES 



119 



was first determined. If a State ranked in the upper 25 percent of this 
range, it was represented by a white oblong, if in the next 25 percent, 
by a light diagonal shading, if in the third, by a darker shading, and 
if in the lowest group, by a solid black oblong. 

This method of classification placed Washington at the head of the 
list of States, with only one feature rated below the upper 25 percent. 
Alabama came last, its ratings in all ten features falling within the 
lowest percentage group. 

FREQUENCY-DISTRIBUTION CHARTS 

The planning and drafting of frequency-distribution charts follow 
closely the methods employed in charting time series in curve and 
column designs. The chief difference is that the horizontal scale repre- 
sents categories of size, value, or some other attribute, instead of 
periods of time. If the basic units are too numerous for individual plot- 
ting, they are grouped in class intervals. This method is illustrated by 
Chart 92, in which each unit of the scale represents a class interval of 
five years. 

MILLIONS 

15 



12 



k 



iiiiiiiiiiiijli 



iiiiiiiiii 







AGE 



Under 5- 10- 15- 20- 25- 30- 35- 40- 45- 50- 55- 60- 65- 70- 75- 
5 9 14 19 24 29 34 39 44 49 54 59 64 69 74 Over 



CHART 92. ESTIMATED POPULATION, UNITED STATES, BY AGE GROUPS: 

1945. 



120 



GRAPHIC PRESENTATION SIMPLIFIED 



The curve design is generally employed for frequency distribu- 
tions of long series which require close plotting. The use of columns 
for short series is illustrated in Chart 93. The data are estimates by 
the Census Bureau of the weekly hours worked during the third quar- 
ter of 1945 by employees fourteen years old and over in non-agricul- 
tural occupations. The horizontal scale represents the categories of 
hours worked, and the columns the number of employees in each 
category. 

The zero base should not be omitted from the amount scale in 
charts of this type. The horizontal scale may begin at any point 
selected as the first unit or class interval of the range covered by 
the series. 




$10.74 



\2.8\$ 



( MILLIONS OF EMPLOYEES ) 






Hours 30-39 40-44 45-49 50-54 55-59 60 or 

more 

CHART 93. ESTIMATED WEEKLY HOURS WORKED, NON-AGRICULTURAL 
EMPLOYMENT: THIRD QUARTER, 1945. 

RELATIVE-VARIATION CHARTS 

A type of chart which was used extensively during the last war 
for comparing monthly fluctuations of stocks and procurement of 
equipment and supplies is illustrated in Charts 94 and 95. The grids 



MISCELLANEOUS TYPES 



121 



of the components in Chart 94 are equal in size. In all of them the 
highest amount of the monthly series was plotted at exactly the same 
distance below the upper line of the grid. The amount scales were 
then adjusted to this measure and the remaining points of the curves 
plotted to the adjusted scales. For example, the highest figure of stocks 
in the first item of Chart 94, "Blankets, Wool, OD, M-1934," was 
5,936,000, reached in February. This high point was plotted one 
eighth of an inch below the upper grid line. In the second chart, 
"Cutter, Wire, M-1938," the high point for the year was 85,000, 
which was reached in October. This amount was plotted at the 
same height above the base as the maximum variation in the first 
chart. The same procedure was followed with the other equipment 
items. In all of them the high point of the series occupies the same 
vertical position. 



BLANKET, WOOL, 00, M-1934 



CUTTER, WIRE, M-1938 




CAN,, MEAT 



Mill ions 




KNIFE, M-1926 



Mill ions 




iiiiiiiiiiiiiiliiilliiiiiiliiiiii 



M A 



J A SO NO 



J F 



AM J J AS 



N D 



CHART 94. RELATIVE MONTHLY VARIATION STOCKS, SELECTED EQUIP- 
MENT ITEMS: 1944. FROM Statistical Yearbook, QUARTERMASTER CORPS. 



122 



FLOUR, WHEAT, WHITE 



GRAPHIC PRESENTATION SIMPLIFIED 

PEACHES, CANNED 



500 


77T? 




- 


100 




Tr^r 


_ 


- 


400 


W.-.Y/ 




- 




















75 


- 






- 


300 


11 




^ 




















Him 






50 


- 








Xvjvi 


200 


Us 


' ;;:::: 
















: : ' : : : : : :v 














100 


?f~:.: 






















81 


25 


-v.Xv 











11 






































Xl3 , , P-JB 











HASH, MEAT AND VEGETABLE, CANNED 



PINEAPPLE, CANNED 



15 - 



10 




60 - 



20 



J 



JF MAM JJAS ON D 



JFM AMJ JA SOND 



CHART 95. RELATIVE MONTHLY VARIATION NON-PERISHABLE SUBSIS- 
TENCE PLACED ON CONTRACT, SELECTED ITEMS: 1944. FROM Statistical 

Yearbook, QUARTERMASTER CORPS. 

Chart 95 shows the same type of comparison in a column design. 
It represents the quantity, in millions of pounds, of various non-perish- 
able subsistence items placed on contract each month in 1943. The 
highest columns in all the charts are exactly the same height, although 
the quantities they represent range from 13 million to 479 million 
pounds. 

These examples were copied from printed sheets 7/2-by-lOM inches 
in size, each containing twelve charts. They are quite as effective for 
comparing relative changes in series of different magnitudes as the 
rate-of-change type of design and are more easily understood. Com- 
pared with the employment of percentages or index numbers, this 
method has the advantage that it obviates the necessity for conversion 
of the basic data to relative numbers. 



MISCELLANEOUS TYPES 



123 



ORGANIZATION CHARTS 

Distributions of functions and personnel may be shown graphically 
in a wide variety of designs. Chart 96, representing a theoretical dis- 
tribution of executive and administrative functions in an industrial 
corporation, is an example of the type frequently employed for this 
purpose. 

The chart units may be represented by oblongs or circles. Oblongs 
are generally preferred if names of personnel are to be lettered within 
the units. Colored inks may be used to differentiate various types of 
functions, the units grading from a strong color for the higher execu- 
tive positions to lighter tints or colors for the minor units. 




CHART 96. TYPICAL ORGANIZATION-CHART DESIGN. 
COMPARISON OF EQUAL AMOUNTS WITH CIRCLES 

Chart 97 is an example of the employment of circles for the com- 



124 



GRAPHIC PRESENTATION SIMPLIFIED 



THE DEFICIT IN PERSPECTIVE 

COMPARISON OF FEDERAL GOVERNMENT EXPENDITURES 
AND DEFICITS, FISCAL YEARS 



DEFICIT 

2 MONTHS, JULY AND AUGUST, 1936 



EXPENDITURES 

i YEAR, AVERAGE, 1900-1913 




DEFICIT 

i YEAR, ESTIMATE, 1939 




DEFICIT 

9 YEARS, TOTAL, 1931-1939 



FROM 178 


9 TO 1642, FISCAL YEARS 


ENDED ON DECE 


MBER 31; FROM 18*4 TO 


OATE.ON JUNE 


30. 


SOURCE ' 


UNITED STATES TREASURY 


DEPARTMENT. 






National Industrial Conference Board 



CHART 97. 



MISCELLANEOUS TYPES 



125 



parison of amounts which are approximately equal in size but which 
differ inherently in a given attribute. The upper and lower circles in 
the first chart represent respectively the Federal deficit for two months 
in 1938 and the average yearly expenditures from 1900 to 1913. The 
second chart is a similar comparison of the estimated deficit in 1939 
with the total expenditures over the 76-year period from 1789 to 1864. 
The third chart compares the total deficit for the nine years from 1931 
to 1939 with the total expenditures over a period of 125 years, 1789 to 
1913. 

CIRCULAR-BAR DESIGN 

In Chart 98 a circular arrangement of the bars representing the 
values is substituted for the vertical sequence usually employed in bar 




CHART 98. EXPORTS, U.S. MERCHANDISE, AS PERCENT OF TOTAL EXPORT- 
ABLE GOODS: 1909 TO 1945. 



126 GRAPHIC PRESENTATION SIMPLIFIED 

charts. The percent figures lettered on each bar show the proportion 
of total exportable goods represented by American exports of mer- 
chandise at various periods from 1909 to 1945. 

Apart from its pictorial quality, this design offers little if any advan- 
tage to the reader over the conventional bar arrangement. It is not as 
easy to read and understand. However, striking variations from the stan- 
dard forms, of this or other types, are sometimes desirable in charts 
for publication in newspapers and popular magazines, where novelty 
in design is required to catch the attention of the casual reader. They 
should be limited to a single aspect of the data, and are usually not 
adapted for presentations involving the use of more than one pattern 
of shading or color. 



.XII 



ELEMENTARY TRAINING METHODS 



ft 



i HE preceding chapters describe 
the main types of design employed in statistical charts. The remain- 
ing chapters are devoted to suggestions and instruction for readers who 
may be interested in preparation for work of this character. 

The simple course outlined in this and the following two chapters 
was planned for beginners who have had no experience in designing 
or drafting statistical charts. Special emphasis is placed on study and 
practice in planning, lay-out, and plotting methods. Manual skill in the 
use of drafting instruments, although essential, is secondary to a thor- 
ough understanding of these fundamental procedures. The use of ink 
should be postponed until a fair degree of proficiency is reached in 
designing and plotting charts in pencil. 

A start can be made with no other equipment than a supply of 
printed graph paper, an ordinary ruler, and a typewriter. Practice with 
ruled forms affords a means for acquiring some knowledge of elemen- 
tary principles without the complication of handling unfamiliar draft- 
ing instruments at the same time. 

127 



128 



GRAPHIC PRESENTATION SIMPLIFIED 



CHARTS ON PRINTED GRAPH PAPER 

Printed graph paper can be obtained in a variety of rulings. Samples 
of five different rulings are shown in Figure 3. Of these, the most use- 
ful for chart purposes are the 5 and 10 per inch. The 6 ruling corre- 
sponds to vertical typewriter spaces, and the 12 and 10 to horizontal 
spaces of elite and pica type. 

In Chart 99 the grid ruling and curve plotting were done before 
the typing of the title and scales. The time scale covers 12 years, but, 
as the plotting is to line, the grid lay-out covers only 11 spaces. The 
ruling of the graph paper is 5 to the inch. Allowing 2 rulings per space, 
the grid measures 22 fifths or 4% inches wide. The amount scale, at 
1 vertical space per billion dollars, takes almost 13 spaces, but, in order 
to keep the scale in multiples of 5 and leave some space above the 
highest plotting, it was extended to 15 fifths or 3 inches. The curve and 
grid lines were drawn with ruler and pencil. 




CHART 99. 



ELEMENTARY TRAINING METHODS 



129 



SUBDIVISIONS 
PER INCH 



6 



8 



10 



12 



Codex Company. Inc. 

FIGURE 3. SAMPLES OF GRAPH-PAPER RULING. 



130 



GRAPHIC PRESENTATION SIMPLIFIED 



After the grid ruling and curve plotting were finished, the title and 
scales were lettered on the typewriter. Special care must be taken with 
typewritten lettering to secure exact centering of time- and amount- 
scale numbers to the corresponding rulings. 

In multicurve charts effective contrast can be obtained by using dif- 
ferent colored pencils for drafting the curves. Graph paper in light-blue 

EMPLOYMENT OF WOMEN 



SELECTED DURABLE GOODS INDUSTRIES 



0000 CCT 



Percent of Wage-Earners 
10 20 30 



1*0 



IRON & STEEL & 
THEIR PRODUCTS 



7.1 0000000 

11.2 mm- 



ELECTRICAL 
MACHINERY 



MACHINERY 5.6 

(EXCEPT ELECTRICAL) 93 



TRANSPORTATION EQUIP. 1,1 
(EXCEPT AUTOMOBILES) 10.5 



NONFERROUS METALS & 172 
THEIR PRODUCTS 16.5 




LUMBER & TIMBER 
BASIC PRODUCTS 



1.7 

5.2 



FURNITURE & FINISHED 9.1 1)0000 
LUMBER PRODUCTS ik.k $m 



ALL DURABLE GOODS 
INDUSTRIES 



War Manpower Commission 



CHART 100. 



ELEMENTARY TRAINING METHODS 131 

ruling is best if the charts are intended for photostat or printed repro- 
duction. The ruled background in this color will not show in photostat 
negatives or printing plates. 

Drafting of bar and column charts on graph paper is equally simple. 
With the variety of rulings available, any combination of bars and 
spaces can be worked out with little difficulty. 

TYPEWRITTEN CHARTS 

Chart 100 is an example of a bar chart done entirely on the typewriter. 
All of this chart was made without removing the paper from the machine. 
The vertical spacing is in single and double space. Each bar is plotted 
by striking the appropriate key as many times as there are units in the 
figure at the left of the bar. 

The title is in spaced capitals underlined, the subtitle in underlined 
capitals. The key for identifying the symbols in the bars is placed imme- 
diately below the subtitle. The upper bar of each pair is made of super- 
imposed parentheses, and the lower bar of the number sign. The dollar 
sign is sometimes used in typewritten bar charts, if the values are in 
money. 

The bars represent rounded numbers. Rounding is necessary in type- 
written charts, as each symbol stands for a complete unit. In the item 
"Electrical Machinery," for example, the 1941 figure, 30.1, is rounded 
to 30 symbols and the 1942 figure, 33.7, raised to 34. 

Column and curve designs take more time and care than the bar 
type. The vertical plotting required in typed column charts involves 
backspacing for each unit in the columns. Curve charts present dif- 
ficult problems in plotting, and the vertical ruling for the grids, 
unless the charts are very small, can be done only on a long-carriage 
machine. 



X 



SMALL CHARTS ON UNRULED PAPER 




HE next step involves the draft- 
ing of small charts on unruled paper, without the aid of typewriter 
spacing or printed ruling. A thorough understanding of the methods 
employed in planning, grid ruling, and plotting of such charts is of 
the greatest importance, as they will be used constantly in all future 
work. 

EQUIPMENT 

The additional equipment needed comprises a drawing board, a 
T-square, two triangles, and a ruling pen. A drawing board about 
12-by-17 inches will be large enough. The wood should be of good 
quality. It is important that the ends and sides be exactly parallel. The 
T-square should be 15 to 18 inches long, with transparent edges. Two 
triangles of transparent material, one 4" 90 and one 6" 30, will be 
required. A small French curve for drawing curved lines, although 
seldom used, may be added. Larger equipment will be needed for 

132 



SMALL CHARTS ON UNRULED PAPER 



133 



work on charts of display size, but is awkward and inconvenient to 
handle in drafting small charts. 

GRID LAY-OUT 

The instructions and examples that follow apply to the lay-out in 
pencil, which should be done before the drafting in ink is begun. 

The initial problem, and the one which usually stumps the begin- 
ner, is how to lay out the grid ruling for the scales in curve charts 
and the bars and spacing in bar and column charts. Assume, as 
a simple example, that the grid of a curve chart 5 inches wide is 




SON 



CHART 101. USE OF SCALE IN GRID LAY-OUT. 



134 GRAPHIC PRESENTATION SIMPLIFIED 

to be ruled for a time scale of 12 months, thus requiring 12 spaces 
for the curve plotting. How is the width of the spaces to be deter- 
mined? 

One can, of course, divide 5 by 12, but the quotient, 0.41666 and 
so on to infinity, is not a practicable unit for drafting. A little elemen- 
tary geometry will solve the problem quickly and more accurately 
than any amount of arithmetical calculation. Place diagonally across 
the grid any measure containing 12 equal spaces that is slightly longer 
than 5 inches. Adjust the angle until the first and last lines or gradua- 
tions coincide with the outside vertical lines of the grid. Mark the sub- 
divisions with pencil dots and rule with T-square or triangle. The 
method is illustrated in Chart 101, page 133. Any measure with equal 
spacing will serve, ruled correspondence paper, an ordinary foot rule, or 
printed graph paper. 

TYPES OF SCALES 

The triangular decimal scale shown in Figure 4 is commonly used 
by draftsmen for lay-out and plotting. It has six scales, with graduations 
ranging from 10 to 60 to the inch. Usually one of these will fit the data. 
If not, multiples of the required number of spaces on one of the scales 
can be used. 




Keuffel and Esser Co. 

FIGURE 4. TRIANGULAR DECIMAL SCALE. 

Lacking a scale of this type, a series of scales can be made from 
Figure 5 that will serve the purpose just as well or better. Each of 
these scales is six inches long. The first has 10 graduations to the 
inch and the succeeding three, 20, 30, and 40. The scales can be 
adapted for use in drafting and lay-out by copying the figures and 
graduations on tapes of tough paper or light cardboard, extending the 
length to about ten inches. This length will be sufficient until work on 
display charts is undertaken, when tapes two or three feet long may 
be needed. 

In some respects tapes have been found to be much more convenient 












= 


ro 








E 


OJ 

^ 








3 























<n = 








Oi 


OB = 








1 
























5 = 








o 




h 




M 


- 








O 


<n z 


OJ 




_ 







* = 
oi = 








"1 


^ ii 








_5 


""1 




CM 




- 


<5 = 
ro = 




> 




* ~ 


o = 

ro = 

oi iz: 

ro ==. 

* 




5 = 
63 ' 




o _ 

M 



135 



136 



GRAPHIC PRESENTATION SIMPLIFIED 



than the triangular scale. One objection to the latter is that it will not 
stay put. When used in scaling or plotting, it must be held in the 
required position, as it is apt to slide about on the chart, particularly 
if the drawing board is set at a convenient slant for drafting. With 
tapes, however, this tendency is obviated by securing each end of the 
scale with adhesive tape. This tape is backed with a light adhesive 
which permits removal of the scale without defacing the paper. It may 
be used also instead of thumb tacks for fastening the paper to the 
drawing board. Another advantage of tapes over scales of wood or 
metal is that they permit free movement of the T-square and triangles 
over the surface of the chart without disturbing the position of the 
scale. 

JOINED-BAR LAY-OUT 

The method for laying out the bars vertically is illustrated in Chart 
102. The chart contains 5 bars. The scale is laid diagonally across the 
grid, with 5 on the upper line of the first bar and zero on the bottom 
line of the fifth bar. The subdivisions are noted from the scale and 
lined with T-square and pencil. 




CHART 102. VERTICAL DISTRIBUTION OF BARS. 



SMALL CHARTS ON UNRULED PAPER 



137 




CHART 103. METHOD FOR PLOTTING AMOUNTS. 

The method to be followed in the plotting of the various amounts 
is pictured in Chart 103. The series comprises the following numbers, 
arranged in descending order: 



Bar 

First 

Second 

Third 

Fourth 

Fifth 



Amount 

645 
520 
423 
255 
146 



The longest bar represents 645. The scale is placed at 7, the equiva- 
lent of 700, on the right edge of the grid, with zero on the base line of 
the bars. The chart shows the use of the triangle for plotting the 
amounts. If a scale is used instead of lettered figures on the bars, it can 
be laid out by the same method employed for plotting the length of 
the bars. 



138 GRAPHIC PRESENTATION SIMPLIFIED 



BAR 



SPACE 



BAR 






CHART 104. BARS AND SPACING IN SPACED-BAR CHARTS. 
SPACED-BAR LAY-OUT 

Chart 104 shows the method for laying out the bars and spaces in a 
spaced-bar design. The spaces are one-half wider than the bars. There 
are 5 bars and 4 spaces, which, at 4 graduations for the bars and 6 for 
the spaces, total 44 graduations. The scale is placed at 44 on the upper 
line of the first bar, with zero on the lower line of the fifth bar. The 
method for plotting the amounts is the same as that used in joined-bar 
charts. 

COLUMN-CHART LAY-OUT 

Chart 105 shows a spaced-column lay-out in which the spaces are 
one half the width of the columns. Each month, with the excep- 
tion of the last, which is not followed by a space, represents 3 
graduations, 2 for the column and 1 for the space, or a total of 35 in 
all. Joined-column charts are laid out in the same way as joined-bar 



SMALL CHARTS ON UNRULED PAPER 



139 




JFMAMJ JASOND 
CHART 105. COLUMNS AND SPACING IN SP AGED-COLUMN CHARTS. 




MAMJJASOND 

CHART 106. PLOTTING OF AMOUNTS ON COLUMN CHARTS. 



140 



GRAPHIC PRESENTATION SIMPLIFIED 




20 



10 



MAMJJASOND 
CHART 107. CURVE-CHART LAY-OUT. 

charts. The only difference is that the scale is horizontal instead of 
vertical. 

Plotting of the amounts is illustrated in Chart 106, on page 139. The 
figure for the highest column is 54. The scale is set at 60 on the upper 
line of the grid to allow room for lettering the figures, if these are placed 
above the columns. In any case some blank space should be left above 
the highest column. This is true also in the case of bar charts and curve 
charts, where some space should be left after the longest bar and above 
the highest point of the curve, unless it is broken. 

CURVE-CHART LAY-OUT AND PLOTTING 

The vertical ruling of the grid in Chart 107 is done in the same 
way as in Chart 101, page 133, except that, as the plotting is to line, 
there are 11 instead of 12 spaces. The curve represents the same series 
of numbers that was used in the spaced-column example and is plotted 
in the same way. 



SMALL CHARTS ON UNRULED PAPER 



141 



946 




523 




CHART 108. CIRCLE- AND SECTOR-CHART LAY-OUT. 
CIRCLE- AND SECTOR-CHART LAY-OUT 

Chart 108 shows a comparison of the numbers 946 and 523 drafted 
as circles proportional in area. The first circle is subdivided in four, 
and the second in three percentage sectors. 

The first step is to determine the diameters of the circles. Assuming 
a diameter of 4 inches for the larger circle, the diameter of the smaller 
one, calculated by the method described in Chapter VII, will be 2.96 
inches (square root of the quotient of 523 divided by 946, multiplied 
by 4 inches ) . A compass will be needed for drawing the circles^, and 
a protractor for pointing off the sectors. (See Figure 6, page 142.) A 
cheap compass, obtainable in any drafting-supply store, will serve for 
pencil sketching. For use with ink a better instrument, with pen attach- 
ment and lengthening bar, will be required. Percentage protractors are 
generally preferred for statistical drafting to those graduated in degrees. 
A numerical series can be more easily converted to percents than to 
degrees of a circle. 

The protractor is laid over the circle and the percents corresponding 
to each sector pointed off on the circumference. The order of plotting 
is by size of sectors, the largest first, the next largest second, and so on. 
Sectors representing a "Miscellaneous" or "All Other" group of com- 



142 



GRAPHIC PRESENTATION SIMPLIFIED 



ponents are placed last in the series of sectors. The plotting begins at 
the center of the upper hemisphere and continues clockwise from right 
to left. 

Familiarity with methods of lay-out and speed in their application 
are the first requirements in the drafting of statistical charts. The 
methods illustrated in the preceding examples are simple, but consid- 





Keuffel and Esser Co. 

FIGURE 6. COMPASS AND PROTRACTOR. 

erable practice will be needed to get them clearly in mind so they can 
be applied easily and quickly. The beginner is advised to repeat the 
examples for the different types a number of times, changing each 
time the dimensions of the grids, the numerical series, and the distribu- 
tion of spacing in bar and column designs, so as to become thoroughly 
familiar with the respective procedures. 



SMALL CHARTS ON UNRULED PAPER 143 



LETTERING 

Typewritten lettering should be used unless a lettering guide is avail- 
able. Hand-lettering on small charts is too slow and usually is less 
satisfactory than typewriting. The machine must give a firm impression 
and good alignment. The large open elite type is preferable to pica and 
larger sizes. Carbon ribbons give clear black copy, but the lettering is 
apt to smudge if the chart undergoes much handling. 

The Vari-Typer machine is used extensively for lettering small 
charts. It has a number of styles and sizes of interchangeable type. The 
quality of the work is superior to typewriting, but operation of the 
machine requires trained personnel. 

If the .chart is intended for printed reproduction, cut-out letters 
can be used for the title. These are backed with an adhesive and can 
be removed and re-used. The titles for Charts 10 and 12 and a num- 
ber of others reproduced in the preceding chapters were made of 
cut-out letters. 




Ralph C. Coxhead Corporation 

FIGURE 7. THE VARI-TYPER. 



PAPER 



Paper for chart purposes should be strong, of uniform thickness, and 
usable on both sides. It can be either smooth or medium, and single 
or double thickness. Bond papers of good quality and smooth finish and 
heavy-weight unruled ledger papers of standard size are extensively 
used. 



144 



GRAPHIC PRESENTATION SIMPLIFIED 



INKING 

After the ruling and plotting are completed, the pencil sketch may 
be lined in India ink with a ruling pen. This instrument is perhaps the 
most useful and the most used in the draftsman's kit. About four fifths 
of all the drafting on statistical charts, outside the lettering, is done 
with the ruling pen. Skill in its use can be acquired only through prac- 
tice. Some suggestions worth keeping in mind are: that no ink be left 
on the outside of the pen after filling; that the supply of ink in the 
pen be kept uniform; that the edge of the pen be kept exactly parallel 
with the guiding edge; and, of special importance, that the pen be kept 
thoroughly clean. Dirty pens do dirty work, and ink left on the pen 
when it is not in use is likely to corrode the metal. 






Keuffel and Esser Co. 



FIGURE 8. RULING PEN. 



Various types of ruling pens, such as dotting and railroad pens, 
used for special purposes, are seldom needed in the drafting of 
statistical charts. Designs containing a large number of small circles, 
of which the circle map is an example, are usually ineffective for com- 
parisons of size, but, if such designs are required, a small bow pen and 
a drop pen for filling in the circles will be needed. For such close work 
the six-inch compass commonly used for circle charts of display size 
is too large. 

Until such time as the beginner has had sufficient practice in 
methods of lay-out and plotting to apply them accurately and without 
hesitation, the drawing should be completed in pencil and carefully 
checked with the data worksheet before the inking is begun. The pencil 
sketch serves as a guide for the ruling pen, thus leaving the attention 
free for the inking process, where a slip or an error is likely to end in 
a difficult and unsightly erasing job. Mistakes in pencil can be easily 
corrected, but once they are inked in they are frequently irremediable. 
In many cases the entire chart may have to be redone. 



SMALL CHARTS ON UNRULED PAPER 



145 





Keuffel and Esser Co. 

FIGURE 9. Bow AND DROP PENS. 
ORDER OF OPERATIONS 

The following order of operations if faithfully adhered to will save 
time and help to insure accuracy: 

1. Lay out and plot in pencil 

2. Check with worksheet of data 

3. Ink with ruling pen 

4. Clean off with artgum any traces remaining of pencil sketch 

5. Lettering 

6. Check lettering with worksheet 

7. Apply shading or color 

8. Inspect and clean 



XIV 



DISPLAY CHARTS 




ISPLAY charts differ in some 
respects from the smaller sizes in principles of construction. The latter 
are usually read at a distance of around two to four feet. The smallest 
lettering on a display chart must be clearly legible at several times this 
distance. If the chart is for use in a large conference or hearing, all 
details should be easily read at not less than twenty feet from the chart. 
Because of the purposes for which display charts are used, greater 
emphasis on simplicity of design and the elimination of surplus word- 
age is required. Unlike the small chart, which is generally viewed under 
conditions that give opportunity for study, the display chart, if for use 
in meetings and conferences, must get its message across in much less 
time. Titles and stubs should be carefully edited to reduce the amount 
of lettering. Long titles and subtitles occupy too much space and throw 
the chart out of balance. As a rule, not more than two aspects of the 
data can be effectively presented on a single chart. If it is not feasible 
to cover the subject in one chart without crowding, it should be broken 

146 



DISPLAY CHARTS 147 

down into two or more. Three simple charts are better than one com- 
plicated one. 

USE OF COLOR 

One of the most significant developments in design and drafting 
technique is the marked increase in the employment of color in all 
types of statistical charts. Willard Brinton's comment in his Graphic 
Presentation, published in 1939, summarizes the views generally held 
by specialists in chart design: "It is believed that the evidence is con- 
clusive that to get the maximum results in graphic presentation the 
question is not 'Can one afford to use color?' but 'Can one afford to 
omit color?' " 

A major contribution to this development has been the adhesive color 
material introduced a few years ago. Apart from its superiority to 
colored inks, paints, and washes in ease and speed of application, it 
has the advantage that the colors can be quickly removed and replaced, 
if corrections or changes have to be made, without defacing or injuring 
the surface of the chart. 

In addition to its use for filling in surfaces with color or shading, this 
adhesive material may be employed in the form of tapes or narrow 
strips in substitution for inked curve patterns in display charts. The 
tapes come in widths from one-eighth inch to one inch. The half -inch 
and quarter-inch widths are frequently used for outlining regional 
subdivisions on maps. 

The presentation of colored display charts in conferences or meetings 
sometimes results in requests for a number of small photostat or printed 
black-on-white copies. If the colors are in paint or water colors, the 
charts will have to be done over in black for reproduction. If adhesive 
colors are used, they can be removed and replaced with black shadings. 
After the printed or photostat copies are made, these in turn, can be 
taken off and replaced with the original colors. The two operations will 
ordinarily take, on a single chart, no more than an hour or so, against 
probably a day's work to make the whole chart over. 

LAY-OUT AND EQUIPMENT 

The methods and procedures described in the preceding chapter 
with relation to lay-out and plotting of small charts apply equally to 
those of display size. A drafting table, adjustable for height and slant, 



148 



GRAPHIC PRESENTATION SIMPLIFIED 




FIGURE 10. DRAFTING TABLE FITTED WITH PARALLEL RULE. 

will be needed. A desirable addition is the parallel rule illustrated in 
Figure 10. This device automatically maintains the rule at right angles 
to the^ sides of the table. Other items which will be required are larger 
triangles, longer scales or tapes, and larger ruling and lettering pens. 

Display charts are generally drafted on illustration board from 20-by- 
30 to 30-by-40 inches in size. The stock should be heavy enough so 
that it will not curl or bend. Highly glazed surfaces which reflect the 
light should not be used. Samples of different boards should be tested 
to see whether they are suitable for drafting. Some are finished with a 
chalky substance that clogs the pens, and in others the direction of the 
grain is a source of difficulty in drafting. 

LETTERING 

Lettering on display charts is usually done with lettering guides. 
Figure 11 gives some idea of the design and method of operation of a 
well-known type. In operation the pen in the forward arm of the scriber 
is guided by the template grooved with the letters and figures. The 
height of the lettering ranges from .08 to 2 inches. 

If equipment for lettering with guides is lacking, the beginner may 
have to start with hand-lettering, although if he remains in this type of 



DISPLAY CHARTS 



149 




Keuffel and Esser Co. 



FIGURE 11. LEROY LETTERING GUIDE. 



work it is likely that he will eventually graduate to the use of guides. 
For this reason it will be advisable to practice some of the styles of 
lettering commonly employed on display charts lettered with guides. 
Early efforts had best be limited to capitals. Small letters are more 
difficult, and in the majority of display charts only capitals are used. 
A sample alphabet is shown in Figure 12, page 150. 

The relation of width to height and the spacing between letters are 
the chief problems 'in hand-lettering. All lettering should be laid out 
in pencil before inking. Large letters may be outlined in pencil with 
stencils or cut-out letters and afterward filled in with ink. In this way 
uniformity is secured, with less chance of the slips and mistakes which 
are likely to occur if the draftsman depends entirely on his own un- 
aided judgment and skill. Little time or practice should be devoted to 
ornate and fancy types of lettering. They are sometimes used in statis- 
tical charts to emphasize a word or words in the title, but the same 
effect can be obtained by underlining. 

PAINTS 

Poster paints are the best color medium for paper or cloth rolled 
charts. The products of different manufacturers vary to some extent in 
adaptability for this purpose, and samples of the various makes should 



ISO GRAPHIC PRESENTATION SIMPLIFIED 

ABCDEFGH 
IJKLMNOP 
QRSTUVW 
XYZ 

abcdefghijk 

Imnopqrstu 

vwxyz 

01234567 
89 

FIGURE 12. SAMPLE LETTERING-GUIDE ALPHABET. 



DISPLAY CHARTS 151 

be tested to insure that one is chosen which will leave a brilliant, 
smooth surface free from brush marks. For stiff surfaces, as has been 
pointed out, the adhesive colors are far superior in appearance and 
are much easier to apply. 

CRAYONS 

Crayons are little used for filling in surfaces on illustration board. 
They are easier to use than paints on rolled charts but do not afford as 
striking contrasts. They are applied with light parallel strokes. The 
color may be deepened by repeating the application. The lines left by 
the strokes of the crayon may be smoothed by rubbing the surface 
lightly with a cloth. 

AIRBRUSH COLORING 

The airbrush can be employed effectively in tinting large maps, but 
is seldom used for coloring statistical display charts. All parts of the 
chart not to be colored must be masked out with paper, and this must 
be secured with adhesive pressed down firmly on the chart to prevent 
the color from creeping under the edges. After the color is applied the 
masking is removed. The whole operation takes much longer than is 
usually required with adhesive colors or paints. If the chart is in several 
colors, the process becomes extremely complicated and time-consuming. 

ROLLED CHARTS 

Display charts light in weight which can be rolled tightly to occupy 
a small amount of space (for shipment by mail or express or to be 
carried as luggage) can be made of tough paper or of cloth similar 
to that used for window shades. Color must be applied with ink, cray- 
ons, paints, or water-colors, as adhesive shadings become detached if 
charts are rolled. 

Rolled charts should be stiffened at top and bottom with light strips 
of wood ( "half-rounds," which can be obtained in any carpenter shop, 
are about the right weight), as illustrated in Figure 13, page 152, to 
prevent the chart from buckling or remaining partly unrolled when in 
use. The strips can be nailed on with short brads. For convenience in 
hanging, a short piece of cord or tape should be fastened to the upper 
strip. 



152 



GRAPHIC PRESENTATION SIMPLIFIED 




FIGURE 13. METHOD FOR STIFFENING ROLLED CHARTS. 
MOUNTING ON CARDBOARD 

Mounting paper charts on cardboard presents no problem where 
facilities for hot mounting under pressure are available, but these will 
be found only in large drafting installations. Mounting by hand is 
usually done with liquid rubber cement. This material will not shrink 
or warp the paper. It is sufficiently strong if a permanent mount is 
not required but loses its adhesive qualities with age. The cement 
is spread uniformly over the back of the chart and the paper pressed 
down firmly to make close contact with the cardboard. The chart should 
be smoothed carefully before the cement dries, to prevent wrinkling of 
the paper. 



XV 



REPRODUCTION PROCESSES 




I HE choice among the various 
methods employed for the reproduction of statistical charts depends 
mainly on the number of copies required. 

PHOTOSTAT 

If the number of copies needed does not exceed fifty, the photostat 
process is generally used. In this process the chart is copied photo- 
graphically on sensitized paper. A copy in white-on-black, called a 
"negative," is made, from which any number of black-on-white copies, 
called "positives," can be duplicated. Figure 14 pictures a half "nega- 
tive" and half "positive" photostatic copy. 

Copies made by this process can be obtained quickly and at moderate 
cost. The standard equipment will make enlargements up to 18-by-24 
inches. If larger sizes are required, the prints may be made in sections 
and joined with adhesive. Charts in color cannot be reproduced by 
the photostat process. 

153 



154 



GRAPHIC PRESENTATION SIMPLIFIED 



EDITORIAL OPTIMISM 



* ABOUT THE WAR 


M 
tr 


arshall f & i 
London Corr* 

I 


r 

gldor / 
1 / 


acArthur 


/ Air Raid 
/ on Japan 


German 
Of f ens 1 ve 


Australia 

<y 


f 




1942 







APR 



MAY 



FIGURE 14. COMPARISON OF NEGATIVE AND POSITIVE PHOTOSTAT PRINTS. 
FROM Manual of Statistics, WAR DEPARTMENT, SOS. 

PHOTO-OFFSET 

In offset printing a sensitized zinc or aluminum plate is made from a 
photographic negative of the chart. On the press the ink adheres only 
to those parts of the plate that contain the image of the chart. In the 
printing process the inked image is transferred to a rubber roller and 
from this to the paper. The average plate can be used to make up to 
twenty-five or thirty thousand copies. 

As in all work for printed reproduction, special care should be taken 
in the preparation of the original charts, for the reason that any imper- 
fections, unless discovered on the negative, where they can be touched 
up, will be reproduced on the plate. Lithographers' artists are skilful 
in patching up defects on plates, but patching of this kind is expensive 
in cost and time. 

Excellent printing in from one to four colors can be done by this 
process, but, as each color requires a separate run through the press, 
color printing is much more expensive than black alone. Some of the 
best published work in colored statistical charts is printed by photo- 
offset. 



REPRODUCTION PROCESSES 155 



MULTILITH 

In this process the chart is drawn directly on a thin aluminum plate. 
It saves the time required for the photographic process and eliminates 
the costly equipment for making photo-offset plates. From six to ten 
thousand copies can be obtained from a single plate. The image is 
usually not as clear as in printing done by photo-offset. 

STENCIL AND GELATIN DUPLICATORS 

For chart production with mimeograph stencils a special drawing 
board called the Mimeoscope is used together with various instru- 
ments designed for drafting on stencils. In reproduction by gelatin 
duplicators the original is typed or drawn with a methyl-violet dye 
and then transferred by contact, in reverse, to a strip of gelatin material 
from which the printing is done. 

These processes are convenient and efficient for duplication of re- 
ports, memoranda, releases, and tabulating forms, but are not very suit- 
able for satisfactory chart reproduction. The results obtained with them 
are not comparable in appearance with those secured with the photostat 
process and with multilith or photo-offset printing. 

ENLARGEMENTS 

Enlargement of small charts for printing is seldom done. Typewritten 
lettering if enlarged over 50 percent generally looks spotty and ragged. 
Slight differences in weight of lettering, due to uneven pressure in 
striking the keys of the typewriter, become prominently visible in the 
enlargement. Insignificant defects in the drawing that would not be 
noticeable in the original stand out clearly. 

Good enlargements to display size of carefully executed originals can 
be obtained by photography, but the cost is usually prohibitive. En- 
largements for display purposes made with the photostat cost less, but 
are much inferior in appearance. 

REDUCTIONS 

Charts for printed reproduction should be made large enough for a 
reduction of 20 to 50 percent. The reduction gives the chart a more 



156 GRAPHIC PRESENTATION SIMPLIFIED 

finished appearance and eliminates minor irregularities in the drafting. 
The amount of reduction a chart will stand is governed chiefly by the 
size of the lettering. Figures 15a and 15b show various types of letter- 
ing and weights of ruling reduced one third and one half of the original 
size. Fine ruling is apt to show breaks or fade out entirely in reductions 



LETTER SIZES 

Slite Type - 12 characters per inch 
Pica Type - 10 characters per 
SMALL GOTHIC - 9 CHARACTER 

LARGE GOTHIC - 9 CHARACTER 
.120" TEMPLATE LETTERING 
.140" TEMPLATE LETTERING 

.175" TEMPLATE LETT 

.240" TEMPLATE 



LINE WEIGHTS 

4 POINT ^-^^ 

3 POINT 

2'/2 POINT 

2 POINT 

I !fe POINT 

I POINT 

3/ 4 POINT 

1/2 POINT 



Original Size 

Note: A point, in printer's measure, is approximately 1/12 of a 
pica, which, in turn is 1/6 of an inch. Therefore, a printer's point 
is approximately 1/72 inch. 



REPRODUCTION PROCESSES 



157 



of more than 50 percent. Reductions of elite typewritten lettering should 
not exceed 30 percent for printing on smooth paper or 20 percent for 
reproduction in newspapers. 

Reduction is calculated in terms of the linear dimensions of the 
chart, not of its area. A 50 percent reduction of a chart 10 inches wide 
by 14 inches long means that the width is reduced to 5 inches and the 
length to 7 inches. The reduction in area, of course, is much greater. 

In practice, instructions to the printer for reduction or enlargement 
are not expressed in fractions or percents. All that is necessary is to indi- 
cate on a line drawn parallel with one side of the chart the exact 
measure to which it is to be reduced or enlarged. 

A problem which frequently arises is to fit a chart into a given space 
in which the proportions of width and length differ from those of the 
chart. For example, suppose that a chart 8-by-15/2 inches is to be re- 



LETTER SIZES 

Elite Type - 12 characters per inch 
Pica Type - 10 characters per 
SMALL GOTHIC - 9 CHARACTER 

LARGE GOTHIC - 9 CHARACTER 
.120" TEMPLATE LETTERING 
.I4CTTEMPLATE LETTERING 

.175" TEMPLATE LET1 
.240" TEMPLATE 



LINE WEIGHTS 

4 POINT 

3 POINT ^ 

2^ POINT 

2 POINT 
I '/ 2 POINT - 
I POINT - 

3 POINT - 



! / 2 POINT 



LETTER SIZES 

Klite Type - 12 character* pr inch 
Pica Type - 10 characters per 
SMALL GOTHIC - 9 CHARACTER 

LARGE GOTHIC - 9 CHARACTER 
.120" TEMPLATE LETTERING 
.140" TEMPLATE LETTERING 

.175" TEMPLATE LET! 
.240" TEMPLATE 



LINE WEIGHTS 

4 POINT 

3 POINT ^ 

2'>*j POINT 

2 POINT 

I '/ 2 POINT 

I POINT 

3^ POINT 

1/ POINT 



Reduced to two thirds of original 



Reduced to one half of original 

FIGURE 15A (OPPOSITE PAGE) AND FIGURE 15e (ABOVE). REDUCTION OF 
LETTERING AND RULING. FROM Time Series Charts, A Manual of Design 

and Construction. 



158 



GRAPHIC PRESENTATION SIMPLIFIED 



ESSENTIAL ITEMS 
Mask, Gaa, Training 
Mine, Land, Chemical 



3,053,190 
1,109,107 



Same 
Size 



ESSENTIAL ITEMS 
Mask, Gaa, Training 
Mine, Land, Chemical 



3,053,190 
1,109,10? 



Reduction 



ESSENTIAL ITEMS 
Maak, Gas, Training 
Mine, Land, Chemical 



3,053,190 
1,109,10? 



20% 
Reduction 



ESSENTIAL ITEMS 
Mask, Gaa, Training 
Mine, Land, Chemical 



3,053,190 
1,109,10? 



30$ 
Reduction 



ESSENTIAL ITEMS 



Mask, Gas, Training 
Mine, Land, Chemical 



3,053,190 40% 

1,109,10? | Reduction 



ESSEHTIAL ITEMS 
Mask, Gaa, Training 
Mine, Land, Chemical 



5,053,190 
1,109,107 



50* 
Reduction 



FIGURE 16. REDUCTION OF ELITE TYPE. FROM Manual of Statistics, 
WAR DEPARTMENT, SOS. 



REPRODUCTION PROCESSES 159 

duced for printing to fit a space 4-by-7 inches in size. The reduction in 
the width dimension, from 8 to 4, is 50 percent, but 50 percent of 15/2 
is 7%, which exceeds the height of the printing space. The reduction 
must, therefore, be calculated from the longer dimension, that is, from 
15M to 7, which amounts to approximately 55 percent. Applying this 
ratio to both dimensions of the chart gives the size of the printing 
image, 3.6 by 7 inches. 

A slide rule is a great convenience for solving quickly and accurately 
problems of this kind. 'If the two chart dimensions are set opposite 
each other on the rule, the conversion to fit a smaller or larger space can 
be determined without computation. 

Figure 16 shows reductions of elite type ranging from 10 to 50 percent 
of the original size. 



XVI 



CLASSIFICATION OF CHARTS BY USE 



IP 

CJftE 



iHE details of the designing, 
planning, and drafting of a chart are necessarily subordinated to the pur- 
pose for which it is to be used. The size of the audience, the physical 
conditions of the locale where the chart is to be shown, and the method 
of presentation are all factors which enter into the problem. Classified 
by use, the majority of statistical charts fall within one of the following 
seven categories: 

1. For publication in newspapers, weekly and monthly magazines, 
or printed reports 

2. For general inter-office distribution 

3. A single copy for desk use 

4. For use in conferences, board meetings, and hearings 

5. For permanent or temporary exhibits 

6. As graphic illustrations to accompany public addresses 

7. For the daily, weekly, or monthly recording of current operations 

160 



CLASSIFICATION OF CHARTS BY USE 161 

CHARTS FOR PUBLICATION 

9 

Statistical charts appear frequently in the reports, bulletins, and 
other published material issued by governmental departments and 
agencies and in trade journals and other periodicals which specialize in 
studies and articles relating to commercial, industrial, or general eco- 
nomic subjects. They are rarely used by the widely circulated popular 
weeklies and monthlies. Occasionally special articles in Sunday news- 
paper editions are illustrated with charts. Some newspapers run line or 
column charts picturing the movements of stock and bond prices. With 
these exceptions statistical charts find little favor in the daily press. 
The average newspaper reader will skip chart illustrations unless they 
are strikingly pictorial or relate to a subject in which he is especially 
interested. The chart has to compete with a variety of other pictorial 
forms which present information, in such a way that it can be readily 
assimilated without mental effort. A statistical chart, no matter how 
clear and simple in design, does demand a little close attention if its 
message is to be understood. 

Bar and column designs are generally preferred to the curve types. 
Sector charts, if they do not contain very small sectors and are not 
overloaded with lettering, are favored for their pictorial quality. Sim- 
plicity in design and lettering is, of course, essential. 

Charts for weekly and monthly magazines require more care in 
preparation than for reproduction in newsprint. Defects that would 
not be noticeable in a newspaper stand out clearly on the smooth-finish 
paper on which magazines are printed. The slight irregularities that 
nearly always occur in typed lettering show up prominently. The grid 
ruling and curves, if curve designs are employed, should be lighter 
than on charts for publication in newspapers. 

The lettering should be done with a lettering guide or typed. Free- 
hand lettering on small charts takes too much time and is usually less 
satisfactory than typewriting. Typed lettering must be clean and neat 
and free from marks of erasures. If the lettering is stripped on the chart, 
the work can be speeded by having the lettering and drafting done 
separately at the same time. 

The drafting and lettering on charts for printed reproduction can be 
done piecemeal. The title, scales, and labels may be lettered on sepa- 
rate pieces of paper and stripped on the chart with rubber cement. In 
this way if a mistake occurs it can be stripped over with corrected 



162 



GRAPHIC PRESENTATION SIMPLIFIED 



copy. Any part of the lettering can be changed as often as may be 
required. The finished chart sometimes looks like patchwork, but if the 
stripping is done carefully, no sign of it will appear in the printed 
copy. The edges of the patches sometimes show a faint dark line, but 
this is easily cleared up by painting with white ink. Chart 109, in which 
the title, subtitle, and curve labels were stripped on after the draw- 
ing was finished, was plated without painting out the edges of the 
stripping. 

WAGE-EARNERS AND PRODUCTION, MANUFACTURING INDUSTRIES 
INDEXES I 1899 TO 1939 

400 



350 



300 



250 



200 



150 



100 




400 



350 



300 



250 



200 



150 



100 



1899 1909 1919 1929 1939 

CHART 109. EXAMPLE OF STRIPPED TITLE AND CURVE LABELS. 

One solution of the lettering problem, illustrated in Chart 110, is 
the substitution of type lettering set by the printer. If this method is 
adopted, the title and labels need not appear on the drawing. An 
additional rough sketch is made for the printer to indicate the location 
and wording of the title and labels to be pasted on the chart before 
the printing plate is made. Charts lettered in this way generally present 
a better appearance than if lettered with guides, and have the advan- 
tage that less time and work in the drafting process is required. 

The use of colors, with the exception of red, which comes out black, 
is inadvisable unless provision is made for color printing. Blues and 
greens come out fainter than red. Reproduction of light tints is usually 



CLASSIFICATION OF CHARTS BY USE 163 



25 



D. C. BUILDING BOOM 

Numerous small projects tote up to a dol- 
lar value matching 1941. 



D. C Building Permits 

[ExeladM Nearby Coustlw) 




'53 '54 *5S '5* '57 'S '5t '40 '41 '48 '45 '44 '43 '4 '47 

*i. 

Date W<nhtn*0e Bal Ejrtcrt. Board Or too w^ton *Mt 



CHART 110. 

too weak for good printing. If the chart is to be printed in color, the 
original, drawn and lettered in black, is accompanied by a rough sketch 
in color to indicate to the printer what portions of the chart are in color 
and what colors to use. 

CHARTS FOR INTER-OFFICE DISTRIBUTION 

Statistical charts are frequently utilized in large offices for circulating 
information of general interest among heads of departments and super- 
visory or administrative personnel. If the number of copies required 
does not exceed fifty, they are generally reproduced by the photostat 
process. The methods used in preparing the charts are the same as for 
printed reproduction. Charts for this purpose are often more technical 
in design and nomenclature than would be practicable or advisable for 
general publication. 

If charts for this purpose form part of a typewritten memorandum 
for photostat reproduction, their width should not exceed that of the 
accompanying text. Otherwise, the reduction necessary to bring the 
chart to the required width will also reduce the width of the text. 



164 



GRAPHIC PRESENTATION SIMPLIFIED 



CHARTS FOR DESK USE 

On small charts for desk use, where only a single copy is required, 
stripping cannot be employed. All the lettering must be done on the 
sheet containing the drawing. If possible, the title should be in larger 
lettering than the stub and labels, and done either with a lettering 

QMC MILITARY PERSONNEL 



THOUSAND 
401 



NON -OPERATING PERSONNEL 



Officer replmctment pool* 




ofc-v 

DEC 



QMC non-operating personnel has 
Increased since March - second quar- 
ter average strength was one-third 
larger than first quarter (17,089 and 
22,996, respectively). Unasslgned 
trainees Increased from 67 to 85 per- 
cent of total In "pipeline* between 
1 January and30 June, while personnel 
In T/0 unit training declined from 21 
to 10 percent of the total Officers 
In pools declined from 1,582 to 800 
in sane period - from 9 to 3 percent 
of total* Percent In schools re- 
mained practically the same. 

Practically no change in June in 
the percent*of operating personnel 
physically qualified for overseas 
service. Percent of physically quali- 
fied officer operating personnel with- 
out overseas service declined from 87 
to 85 percent of total qualified in 
June. Corresponding percentage for 
enlisted personnel Increased from 60 
to 64 percent of total in June. 



MALE OPERATING PERSONNEL 



PERCENT QUALIFIED FOR 
OVERSEAS SERVICE 



Officers 



Enlisted 



PERCENT OF PHYSICALLY QUALIFIED 

PERSONNEL WITHOUT OVERSEAS SERVICE 

30 June 1945 





PERCENT 




Officers 



Enlisted 
Men 



PERCENT 



35 and / 
over/ 



CHART 111. FROM STATISTICAL REPORTS TO THE QUARTERMASTER 

GENERAL. 



CLASSIFICATION OF CHARTS. BY USE 165 

guide or typed. Free use of color will aid in setting off contrasting 
elements and contribute to the interest and attractiveness of the chart. 

The lettering on Chart 111, including the title, was all done on 
the Vari-Typer. The variety of styles and sizes of interchangeable 
type provided by this machine is especially advantageous for letter- 
ing small charts. This example was selected from a series of daily re- 
ports to the Quartermaster General. The accompanying text comment 
and summarization of the data was featured in all the reports. 

The objective was to place at the disposition of the Quartermaster 
General as quickly as possible the statistical background relating to 
questions of administration and policy requiring his attention and ac- 
tion. Through contacts with the chief executive officers the statistical 
department was kept in close touch with current developments. The 
selection of subjects for the charts was made on the basis of their 
immediate interest and importance. 

The drafting was done on sheets of standard size, punched for filing 
in loose-leaf binders. Colored shadings were used extensively, but were 
selected from the patterns most suitable for photostat reproduction, 
in case a number of copies were needed. 

Text summaries are a valuable addition to any chart, but, because of 
the prohibitive amount of space they would occupy in the large letter- 
ing used on display charts, their use is generally limited to charts 
of small size. 

Charts of this type fill a real need. Chart and exhibit rooms serve 
very well to show to visitors, but it can hardly be expected that a busy 
executive will depend on them to any great extent for the class of cur- 
rent information he needs to have immediately available in a form that 
facilitates quick reference and of a size convenient for desk use. 

CHARTS FOR CONFERENCES, BOARD MEETINGS, AND HEARINGS 

The selection of design and size of display charts for conferences, 
meetings of boards or committees, and other small gatherings is deter- 
mined mainly by the number of oeople likely to be present. For an 
ordinary departmental conference of executive heads, display charts 
around 20-by-30 inches in size will be large enough. If the number 
taking part in the conference exceeds fifteen or twenty, both lettering 
and drawing should be larger. 

If the conference includes representatives of interests widely diverse 
in character, careful editing of the titles and labels is necessary to make 



166 GRAPHIC PRESENTATION SIMPLIFIED 

them clearly descriptive in terms familiar to everyone. Labels should be 
placed close to the components and categories. Attempts to show dif- 
ferent aspects of the data in a single chart are usually ineffective. Unless 
the chart automatically tells its story it will have to be explained, and 
charts that require verbal explanation are generally failures. Redundant 
lettering, unessential ruling, and ornamental borders and curlicues 
serve only to divert attention from the main purpose of the chart. 

The lettering should be done, if possible, with lettering guides. Hand- 
lettering of display size generally lacks finish and uniform spacing. 
Irregularities in lettering are the first details to catch the eye, and the 
impression of carelessness or poor workmanship they create detracts 
seriously from the effectiveness of the chart. 

The time between the calling of a conference and the date it is to be 
held is sometimes too short for making up the required number of 
charts in display size. A method that may be employed in such emer- 
gencies consists in drafting the charts on small sheets, with stripped 
typed lettering, for duplication by photostat. Enough copies of the 
originals are run off to provide each person attending the meeting with 
a full set of the charts planned for presentation. 

To obtain the best results with charts for use in conferences or 
hearings it is essential that the method of presentation be carefully 
planned and adapted to the order of business likely to be followed dur- 
ing the meeting. It frequently happens in business conferences that no 
formal agenda is prepared and the order in which the various subjects 
will come up for discussion cannot be determined in advance. The best 
plan in such cases is to distribute the charts separately about the room, 
so they are immediately available when .they are reached in the course 
of the discussions. 

If the agenda is limited to one general subject, the procedure may 
be similar to that followed in lectures. A speaker, sometimes the chair- 
man of the conference or a member of the statistical staff that designed 
the charts, takes charge of the presentation. The charts, placed on an 
easel or a table, are shown and discussed one at a time. 

Various details that must be taken into account have to do with the 
physical conditions of the room in which the meeting is held. They 
usually offer little difficulty if it takes place in an office or board room 
equipped with facilities for the placement and handling of the charts. 
But if it is scheduled for a locale where such essentials as adequate 
lighting and the necessary furniture and equipment for displaying the 
charts are lacking, these must be provided before the meeting begins. 



CLASSIFICATION OF CHARTS BY USE 167 

Any last-minute improvisations are very likely to prove unsatisfactory. 

Nothing so tries the patience of an audience as confusion or delays 
in the order and method of the presentation. The position and height 
of the chart stand and adequate lighting are of special importance. 
Some inconspicuous procedure for disposing of the charts after they 
are shown is needed. Charts not exceeding 20-by-30 inches in size 
usually can be laid face down on the table in front of those not yet 
shown. Charts of larger sizes should be turned over to an assistant for 
disposal. The speaker should stand at the right, not in front, of the 
charts, so he will not obstruct the vision of people seated at the side 
of the room. He can use a pointer two or three feet long to call atten- 
tion to features of the chart of special significance, without at any 
time impeding a full view of the chart from all parts of the room. 

A brief reference to the experience of the War Department Statistics 
Branch in the course of the series of meetings with the Senate and 
House Military Committees during the last war, referred to in Chap- 
ter I, may serve to illustrate the type of problems met with in the use 
of charts in conferences. Many of the meetings were held in the offices 
of the War Department, where all necessary equipment and facilities 
were available. Often, however, for the greater convenience of the 
Committee members, they took place in committee rooms in the Capi- 
tol or in the Senate and House Offices, where a number of the re- 
quirements for successful chart presentation were lacking. Also, as the 
rooms were in frequent use for other purposes, the time for remedy- 
ing such deficiencies was limited. 

In the meetings with the Senate Committee where the attendance 
seldom exceeded ten, little difficulty was encountered in securing a 
table or desk large enough and high enough to display the charts full- 
length to all present. The provision of supplemental lighting was usu- 
ally unnecessary. As none of the members of the Committee were more 
than twenty feet from the charts, the labels and titles were clearly 
legible and reference to them could be omitted in the speaker's verbal 
introduction to each chart. 

The attendance in the meetings with the House Committee gener- 
ally ranged from thirty to forty. This meant that if the charts were 
placed on an ordinary desk or table only the upper part of them would 
be visible to those seated beyond the fourth or fifth row of chairs. The 
stand for the charts had to be built up by one means or another 
sometimes by piling books on the table or adding a pair of filing-case 
units to secure sufficient height. Usually the illumination from the 



168 GRAPHIC PRESENTATION SIMPLIFIED 

overhead lighting was inadequate. This was remedied by placing a 
footlight below the charts. As the labels of curves and segments could 
not be easily read from the rear of the room, the speaker included in 
his introductory remarks a complete description of the graphic features 
of each chart. 

Questions from the audience were another serious problem. The 
speaker had to be fairly well informed concerning the subject of each 
chart and have at hand in convenient form such additional data as 
might possibly be requested by members of the Committees. A fre- 
quent request, if the charts represented percentages or index numbers, 
was for their translation into their basic equivalents, or in the case of 
curve charts, the high, low, and end points of the curves. Such ques- 
tions were taken care of by lettering the figures in light pencil above 
or below these points or on the margins of the charts, where, although 
not visible to the audience, they were instantly available in answering 
questions of this kind. Requests for more extensive data were referred 
to an assistant provided with tables and other source material per- 
taining to the charts. 

A few experiments were made with pictorial additions to the charts 
sometimes characterized the "kindergarten" method consisting in 
drawings of the objects represented by the curves, bars, and other con- 
ventional graphic types. For example, if the chart pictured the pro- 
duction rate of airplane bombs, a drawing of a bomb accompanied 
the lettering in the stub in bar charts or above the curve in curve 
charts. It was finally decided that the limited value of such aids to 
visualization did not justify the time and labor expended on them. 

CHART EXHIBITS 

In large offices a room is sometimes set aside to be used exclusively 
for a permanent chart exhibit. In other cases the exhibit is located 
in a room used for conferences and board meetings. 

The commonest defect in chart exhibits is lack of flexibility. The 
exhibit rooms are filled up with subjects which appear to be of out- 
standing interest at the moment. Within a short time many of the 
charts will be out of date unless provision is made for keeping the 
exhibit in line with the new statistical data received from day to day. 
As many as possible of the charts should be designed with space left 
for adding new data. If the time scale in curve and column charts is 
in months, blank space for at least three additional months should be 



CLASSIFICATION OF CHARTS BY USE 169 

left at the right of the drawing. Otherwise, the drafting room will 
be confronted each month with the task of making over most of the 
live charts in the exhibit. 

The major problem in the maintenance of a chart exhibit is how to 
keep it alive. Every facility for making frequent changes should be 
provided. The method employed for fastening the charts in place on 
the walls or display fixtures should permit their easy removal without 
tearing or defacing the paper or cardboard. Stapling devices are not 
adapted for use on charts that must be returned to the drafting room 
periodically to be brought up to date, or that are likely to be needefl 
in conferences or for exhibit in another room or building. Within a 
short time the margins on which the stapling is done become defaced 
to such a degree that the chart has to be made over. Map tacks, which 
leave little trace on the chart margins, are commonly used for this 
purpose. 

The type of fixture which is shown in Figure 17 is very convenient 
for exhibits of small- or medium-sized charts. It occupies little space 
and is portable. The use of wall space is preferable for charts of display 
size. 




Multiplex Display Fixture Company 

FIGURE 17. PORTABLE FIXTURE FOR EXHIBIT OF SMALL CHARTS. 

Conferences requiring the use of charts are frequently held in rooms 
where hanging the charts on hooks or tacks is impracticable because of 
possible injury to the wood paneling or the wall finish. In such cases 



170 GRAPHIC PRESENTATION SIMPLIFIED 

light easels are the best solution, although they have the disadvantage 
that only one chart at a time can be shown. 

One of the most important requirements for the display of statistical 
charts in exhibits is adequate lighting. It must be strong enough to 
bring out clearly contrasts in color and shading and permit easy reading 
of the lettering. Overhead central lighting is usually unsatisfactory. 
Whenever possible, hooded lights should be placed immediately above 
the charts. 

Figure 18 shows a wall cross-section of a permanent chart exhibit 
installed by a Federal agency in Washington. Good lighting is secured 
with a series of fluorescent lights extending the length of the wall. These 
are enclosed in a lighting box open at the bottom. The space between 
the lighting box and the shelf below is large enough for a chart 40 to 
42 inches in height. This space is backed with plywood, with an outside 
layer of heavy cardboard. The slanting shelf below the charts provides 
space for reports, tables, or other matter relating to the subjects 
presented in the exhibit. 

CHARTS FOR PUBLIC ADDRESSES 

The chief problem in the preparation of charts for use with addresses 
or discussions before large gatherings is the matter of size. The lettering 
on a 30-by-40 inch illustration-board chart, unless it is very large, will 
not be clearly legible at a distance beyond thirty to forty feet. Larger 
sizes of illustration board are difficult to transport and awkward to 
handle. 

If the address is given from the seating level of the room, the utiliza- 
tion of illustration-board charts is impracticable. Only those of the 
audience seated near the speaker can obtain a full-length view of 
the charts. If they are shown from a raised platform, the stand for the 
charts should be high enough to leave a space of several feet between 
the lower margins and the floor. 

The charts should be shown from a central position, one at a time. 
Distributing them separately at the front of the stage or platform, 
an arrangement sometimes adopted, saves some handling, but divides 
the attention of the audience, which should be held to the chart under 
discussion by the speaker. 

Strong lighting should be provided if cardboard charts are used, 
otherwise only those of the audience seated near the speaker will be 
able to distinguish details in the charts. Footlights, sometimes used for 



CLASSIFICATION OF CHARTS BY USE 



171 




f LOOU 
FIGURE 18. WALL CROSS-SECTION OF A CHART EXHIBIT. 



172 



GRAPHIC PRESENTATION SIMPLIFIED 



this purpose, are better than nothing, although the lighting should be 
from above rather than from below the chart. 

Lantern-slide projection gives the best results if the audience is 
too large for the use of illustration-board charts. The charts are first 
drafted in a size convenient for lettering and then reduced photo- 
graphically to that of the slides. With a good projector a clear magnifica- 
tion up to ten feet square or over can be secured. 




Golde Mfg. Co. 



FIGURE 19. LANTERN-SLIDE PROJECTOR. 



While the best results in lantern-slide projection will be obtained 
only if the room where the charts are shown is in complete darkness, 
this is often difficult or impossible in large public halls and audi- 
toriums. The disadvantages of semidarkness can be overcome to a 
considerable extent if a projector with a high light output is used and 
the position of the screen with relation to the room lights adjusted so 
that no light except that from the projector falls directly upon it. The 
room lights should be controlled, if possible, from the projector position. 



CLASSIFICATION OF CHARTS BY USE 173 

This method solves the problem of size, but involves a few minor 
complications, among them the position of the speaker. He cannot 
stand in front of the screen, and at either side will not be seen clearly 
in the semidarkness required for satisfactory screen projection. This 
difficulty is usually met by placing in front of the speaker a small 
table or desk equipped with a hooded desk light. 

The microphone, commonly employed for the delivery of addresses 
to large audiences, limits to some extent the speaker's freedom of move- 
ment. Unless he is accustomed to the use of this instrument, he is likely 
in the course of his address to turn his head toward the screen to call 
attention to some especially significant feature of the chart. This will 
be instantly reflected in a noticeable difference in the tone and volume 
of his voice. 

The projector, with the slides arranged in their appropriate order, 
should be in charge of an assistant who can be trusted not to mix up 
the chart sequence nor put the slides into the projector upside down. 
A system of visual signals between the speaker and his assistant should 
be worked out and practiced until there is no chance of slips in the 
timing. 

Charts to be shown with a public address should be extremely 
simple in design. Colors, shading, and curve patterns designating 
components and categories should be chosen with an eye to strong 
contrasts. It is essential that the meaning and purpose of the chart be 
obvious at a glance. If they are not, the attention of the audience is 
divided between deciphering the chart and listening to what the 
speaker has to say about it. 

CHARTS FOR RECORDING CURRENT OPERATIONS 

Charts of this type are little more than accounting devices in graphic 
form. They are used extensively in all classes of industrial and com- 
mercial activities. They generally consist of a series of dates at the top 
of the chart and a list of items at the left. The ruling is vertical for 
the -time scale and horizontal for the items. The entries correspond- 
ing to each time unit are noted as numbers or symbols or with ad- 
hesive tape in color or shading. If the chart is laid out on cardboard, 
map tacks of different colors or shapes may be utilized for the same 
purpose. 

Ruled loose-leaf sheets are commonly employed for charts repre- 
senting a continuous daily record. If the entries are at longer intervals, 



174 GRAPHIC PRESENTATION SIMPLIFIED 

the chart is usually drafted in wall-chart size. A curve or column dis- 
play chart, if space is left at the right of the drawing for future 
entries, serves the same purpose. 

A type sometimes designated "Progress Chart," similar in design, is 
employed to record progress on a project, study, or investigation with 
a definite time limit. Usually the time scale is in months, the last month 
in the series representing the date set for completion. The entries are 
in percents, showing the proportion of the project finished to the end 
of each time unit. 



XVI 



STATISTICAL PROCEDURES 



IP 

ulii 



I HE foregoing chapters have 
dealt chiefly with the various stages of design, drafting, and production. 
This chapter will take up in some detail the procedures employed in 
the preparation of the statistical material before the chart reaches the 
drafting stage. 

In large enterprises or organizations this is a matter with which the 
draftsman need not concern himself. The initial idea of the chart gen- 
erally originates with someone exercising managerial or administrative 
functions. It then passes to the statistical staff, which carries out the 
necessary research and the collection and analysis of the data. It usually 
reaches the draftsman in the form of a table suitably arranged for 
drafting and generally accompanied by a rough freehand sketch out- 
lining the design selected. 

There are many small enterprises, however, where the draftsman may 
be called upon to take over partly or entirely the work of collection and 
analysis of the data and the preparation of the tables for the charts, as 
well as their reproduction in typed or printed form. In this case he 

175 



176 GRAPHIC PRESENTATION SIMPLIFIED 

will need some knowledge of statistical procedures. A large number 
of persons are engaged in such work, but only a small proportion call 
themselves statisticians, and fewer still have had special training in 
statistics. The fact is that the methods employed in the preparation of 
data for the majority of charts are simple and easily learned. Some of 
the most important of these are outlined in the following pages. A brief 
study of a good text on statistics will give the student all the additional 
refinements in statistical technique he will ever be called upon to use. 
Elementary Statistics, by Dr. F. H. Harper, may be especially recom- 
mended for simplicity and clarity of treatment in terms understandable 
to anyone with a high-school education. 

The knowledge of mathematics needed for this phase of the work 
varies with the nature of the problem. Only rarely will more than a 
thorough grounding in arithmetic be required. This is true even of the 
higher flights of statistical analysis, once they are divested of technical 
terms and mathematical symbols. After the various steps are broken 
down into everyday English, it will be found that they demand only a 
working knowledge of eighth-grade arithmetic. 

ROUNDING 

Much time will be saved in preparing statistical data for charts or 
tables by rounding large numbers to three or at the most four digits. 
Plotting to the fourth digit in any but very large display charts is im- 
practicable. In tables small numbers save space and are easier to read 
and understand than large ones. Rounding focuses attention on the 
more significant digits and reduces the work of calculating the conver- 
sion of basic figures to relative numbers. 

In rounding, a digit less than five is dropped. A digit greater than five 
adds one to the next digit to the left. If the digit to be discarded is an 
even five, the next digit, if an even number, is unchanged, but if odd, 
it is raised. 

Care must be used in rounding series of numbers varying widely in 
magnitude to avoid undue distortion of the smaller items. For example, 
the series 150,700, 72,300, and 5,500, if rounded to thousands, becomes 
151, 72, and 6. The change in the first two numbers through rounding 
amounts to less than one half of 1 percent, but the third number is in- 
creased by over 9 percent. It is best in such cases to add one decimal 
place in all the rounded numbers. 

Frequently a column of rounded numbers will not add to the 



STATISTICAL PROCEDURES 177 

rounded total. The difference is usually slight, but a footnote should be 
added explaining that, because of rounding, the total does not agree 
with the sum of the items. If the figures are percents, it is customary to 
increase or decrease by one the right-hand digit of one or more of the 
larger percents to bring the total to 100. 

AVERAGES 

Averages are employed to summarize statistical data. They are com- 
puted in various ways. The terms designating the two methods most 
frequently used are "simple," or "unweighted," and "weighted" aver- 
ages. The simple average is merely the quotient obtained by dividing 
the total of a series of values by the number of components. Table IV 
illustrates the procedure. 

TABLE IV 

SIMPLE AVERAGE 

Components Values 

A 220 

B 135 

C 365 

Total 720 

Average 240 

If the series contains one or more extreme values, the average may 
differ so far from any of the individual values or any grouping of them 
as to be of no real significance. Table V is an example. 

TABLE V 

AVERAGE OF A WIDE RANGE OF VALUES 

Components Values 

A 75 

B 306 

C 34 

D 26 

E 550 

F 65 

G 29 

Total T 1,085 

Average 155 



178 GRAPHIC PRESENTATION SIMPLIFIED 

The figure obtained by dividing the sum of the values by the number 
of components is nearly three-and-one-half times the average of the five 
smaller values and is 64 percent below the average of the two larger 
ones. In cases of such wide differences in the range of values, the 
median, the middle item when the values are arranged according to 
size, is sometimes used. This arrangement, using the same data, is 
shown in Table VI. 

TABLE VI 
ARRANGEMENT OF VALUES TO DETERMINE THE MEDIAN 

550 
306 

75 

65 Median 

34 

29 

26 

The median, 65, is more representative of the five lower values than 
the average, but neither is of much use as a summarization of the whole 
series. The median is seldom employed in commercial and industrial 
tables, chiefly because few people are familiar with the term. The same 
objection applies in lesser degree to "mean" and "arithmetic mean," 
which are sometimes used for "average." 

WEIGHTED AVERAGE 

If both components and values vary in size, the weighted average 
should be used. Table VII shows the weighted-average cost per unit 
of a product manufactured in three separate establishments. 

TABLE VII 

WEIGHTED AVERAGE 

Number of 

Establishment Units Produced Total Cost 

A 400 $ 4,800 

B 1,100 11,900 

C . 700 8,600 

Total 2,200 $25,300 

Weighted-average cost per unit $11.50 



STATISTICAL PROCEDURES 179 

Weighted averages derived from series in which there are extreme 
differences in the size of components or values are likely, as in simple 
averages, to be of little significance for practical use. 

Averages and percents should not be averaged. The procedure for 
obtaining a true average of a number of averages or percents is the 
same as that for the weighted average. The basic values and com- 
ponent figures from which the individual averages or percents are 
derived are totaled. The true average is the quotient of the value total 
divided by the component total. 

A type of error common in both simple and weighted averages is the 
inclusion of components which have no bearing on or merely distort the 
summarization. Errors of this kind are frequent in per capita estimates 
covering the total population. A recently published rate of per capita 
consumption of liquor in the United States was calculated by dividing 
the total consumption by the total population, without taking into ac- 
count that the population group fourteen years of age and under, num- 
bering nearly thirty-three million, consumes no liquor. 

FREQUENCY DISTRIBUTIONS 

Frequency distributions are distributions of amounts by categories. 
Table VIII, representing a theoretical distribution of the wage-earners 
in an industrial establishment by rates of pay per hour, is an example 
of a simple frequency distribution. 

TABLE VIII 

SIMPLE FREQUENCY DISTRIBUTION 
Rates of Pay Number of Wage-earners Re- 
gents Per Hour) ceiving Specified Rates 

70 12 

75 21 

80 63 

85 75 

90 84 

95 80 

100 74 

105 60 

110 51 

115 36 

120 and over 22 



180 GRAPHIC PRESENTATION SIMPLIFIED 

Table IX shows a distribution of wage-earners by weekly earnings, 
estimated from the data in Table VIII. The number of different individ- 
ual amounts in an average weekly payroll is too large to show each sepa- 
rately in a chart or statistical table of reasonable size. For this reason, 
the data are grouped in class intervals. 

TABLE IX 

CLASS-INTERVAL DISTRIBUTION 

Class Interval Number of 

( Dollars Per Week ) Wage-earners 

25-29 31 

30-34 140 

35-39 235 

40-44 114 

45 and over 58 

The class intervals in this table are in groups of five, each interval 
comprising the items from the first to the last number, including 
those exceeding the last, but less than the first number of the 
following class interval. For example, a wage-earner whose pay check 
amounted to over $44, but less than $45, was classed in the 40-44 
interval. 

The class intervals should be approximately equal in size. Frequently 
a series is made up chiefly of items which fall within the range of rela- 
tively small class intervals but ends with several of much greater spread. 
In Table X the range in each of the first four class intervals is 100, while 
in the fifth it is 500, and in the sixth, 1,000. Presented in a table, a dis- 
tribution of this kind may be understood, but it is difficult to make a 
chart of it that will show a clear picture of the relation between the 
amounts in the small and large class intervals. 

TABLE X 

DIFFERENCES IN RANGE OF CLASS INTERVALS 

Class Interval Amount 

101-200 125 

201-300 224 

301-400 654 

401-500 185 

501-1,000 110 

1,001-2,000 56 



STATISTICAL PROCEDURES 181 

INDEX NUMBERS 

Index numbers are frequently used in the analysis and presentation 
of statistical data. Published index-number series compiled by govern- 
mental statistical agencies alone cover an extensive range of economic 
and industrial activities. Many tables published in economic periodicals 
include only index numbers, omitting the basic data from which the 
indexes are derived. 

The term index number has a technical sound, but it is only another 
name for percent. A certain figure or figures in a series is selected as 
the base, with an assumed value of 100. All the figures in the series are 
then converted to percents of this base. For example, the series 160, 80, 
120, 200, if the base selected is the first number, becomes as index num- 
bers 100, 50, 75, 125. Each of the numbers is divided by the base. 

The designation "index number" is frequently omitted in tables and 
charts. The base figure, followed by 100-for example, '1929 = 100," 
"1941-45 = 100" is lettered in the body of the chart or over the upper 
ruling of the grid. In typewritten tables it is either typed below the title 
or over the first ruling of the table. In printed tables it is often set in 
heavier type. 

The examples which follow illustrate some of the advantages of 
this method for use in statistical analysis and particularly its value 
for comparisons of two or more series of unlike basic units or of 
extreme difference in size. As the examples are interrelated, the same 
time series, 1929 to 1937, is used in all of them. This period covers 
the great depression of 1929-33 and the succeeding recovery move- 
ment during which various important indicators reached or exceeded 
the 1929 levels. The data represent important factors in the manufac- 
turing industry; the charts present the same data in graphic form. 

The first example shows the use of index numbers for comparing two 
series which, because of the difference in the nature of the basic units, 
are not comparable in their original form. The series represent two fac- 
tors of primary importance, the total number of workers employed and 
the total amount of time worked. The data in Table XI are estimates of 
average employment and average hours worked per week in the manu- 
facturing industries. The estimates were published by the Bureau of 
Labor Statistics of the U.S. Department of Labor. The employment data 
include both wage and salary employees, but the changes over the 
period may be taken as representative of the employment of wage- 



182 



GRAPHIC PRESENTATION SIMPLIFIED 



earners, as the salaried employees constitute only an insignificant pro- 
portion of the over-all totals. 

TABLE XI 

EMPLOYMENT AND HOURS WORKED PER WEEK, MANUFACTURING INDUSTRY: 

1929 TO 1937 

1929 1930 1931 1932 1933 1934 1935 1936 1937 
10,534 9,401 8,021 6,797 7,258 8,346 8,907 9,653 10,606 

44.2 42.1 40.5 38.3 38.1 34.6 36.6 39.2 38.6 



Year 
Employment 

( Thousands ) 
Hours worked 

per week 



The two series are drafted separately in Chart 112. Double-scale 
charts are sometimes used for comparisons of this kind. Whether com- 
bined in a single chart with two scales or charted separately, however, 
the curves, because of the difference in the scales, do not afford a true 
comparison of the relative changes over the period and offer but little 
advantage over the tabular presentation. 



THOUSANDS 
12 



EMPLOYMENT AND AVERAGE WEEKLY HOURS, WAGE-EARNERS IN MANUFACTURING: 1929 TO 1937 

HOURS 
50 




Employment 



Average Weekly 
Hours 



1929 '30 '31 '32 



35 '36 '37 



1929 '30 '31 '32 '33 '34 



CHART 112. 

Both series, converted to index numbers based on 1929 as 100, with 
a third index, total hours worked, derived from the first two, are shown 
in Table XII and Chart 113. 



STATISTICAL PROCEDURES 



183 



TABLE XII 

EMPLOYMENT, HOURS WORKED PER WEEK, AND TOTAL HOURS WORKED: 

1929 TO 1937 
INDEX NUMBERS, 1929 = 100. 



Year 

Employment 
Hours worked 

per week 
Total hours 

worked 

( Hours worked per 
week multiplied 
by employment ) 



1929 
100 
100 



1930 
89 
95 



1931 
76 
92 



1932 
64 
87 



1933 
69 
86 



1934 
79 
79 



100 84 70 56 59 62 



1935 
84 
83 

70 



1936 
92 
89 

82 



1937 

101 

87 

88 



AVERAGE EMPLOYMENT AND HOURS WORKED 



1929 '30 '31 '32 '33 '34 '35 '36 '37 




184 GRAPHIC PRESENTATION SIMPLIFIED 

Either of two methods may be employed to obtain the third series, 
total hours worked: (1) Multiply average hours worked by number 
employed and convert the products to index numbers; or (2) multiply 
the index of hours by the index of employment and divide the products 
by 100. The latter method is simpler and shorter, and the index num- 
bers, after rounding, will be the same. Index numbers from 100 to 1,000 
are rounded to three digits, and if less than 100, to two digits. Slightly 
closer accuracy can be secured by adding one decimal place, but this 
is offset by loss in simplicity and clarity of presentation. There is no 
point in straining for microscopic accuracy in relative series which, in 
reality, can be only approximately exact, particularly if the data repre- 
sent estimates or averages combined from many sources. 

Conversion of the basic data to index numbers makes it possible to 
present them graphically in a form that facilitates exact comparison and 
interpretation. 

The volume of industrial production is one of the primary indicators 
of changes in economic conditions. The first line in Table XIII repre- 
sents the index for all manufacturing for the years 1929 to 1937 based 
on the average for the period 1935-39. These data were compiled by the 
Federal Reserve Board. The second line is the same index converted to 
a series based on 1929. The conversion is effected by dividing each of 
the numbers in the Federal Reserve Board index by its 1929 index 
number, 110. 

TABLE XIII 

PRODUCTION, WHOLESALE PRICES, AND VALUE OF PRODUCTION: 

1929 TO 1937 

Year 1929 1930 1931 1932 1933 1934 1935 1936 1937 

Industrial production 110 90 75 57 68 74 87 104 113 

(1935-39 = 100) 
Industrial production 100 82 68 52 62 67 79 94 103 

(1929 = 100) 
Wholesale prices 100 93 82 74 75 83 87 87 92 

( Manufacturing ) 

Value of production 100 76 55 39 46 61 69 91 95 
( Price index multi- 
plied by production 
index ) 

The first two lines illustrate the procedure for converting series of 
index numbers from one base to another. The wholesale-price index is 
compiled by the U.S. Department of Labor. The fourth line, value of 



STATISTICAL PROCEDURES 



185 



production, is calculated by the short-cut method described in the pre- 
ceding example. The three series, production, price, and value, are com- 
pared graphically in Chart 114. 

Special care should be exercised to obviate misinterpretation or mis- 
use of derived indexes through the employment of designations that 
imply a relationship which may be, in fact, non-existent. The indexes 
"Output Per Worker" and "Per Hour," derived by dividing the index of 
production by the indexes of employment or man-hours, are examples. 
Changes in these indexes are sometimes assumed to indicate that cor- 
responding changes have taken place in the productiveness or efficiency 



PRODUCTION, WHOLESALE PRICE 
AND VALUE OF PRODUCTION 




186 GRAPHIC PRESENTATION SIMPLIFIED 

of labor, although this interpretation has no basis in fact. The course of 
these indexes through the period 1929 to 1937 illustrates this point. Both 
the basic and the derived indexes are shown in Table XIV. 

TABLE XIV 

OUTPUT PER MAN AND PER HOUR, MANUFACTURING: 1929 TO 1937 

Year 1929 1980 1931 1932 1933 1934 1935 1936 1937 

Basic Indexes 

Employment 100 89 76 64 69 79 84 92 101 

Total hours worked 100 84 70 56 59 62 70 82 88 

Output (Production index) 100 82 68 52 62 67 79 94 103 

Derived Indexes 
Output per man ( Output index 100 92 89 81 90 85 94 102 102 

divided by employment index ) 
Output per hour ( Output index 100 98 97 93 105 108 113 115 117 

divided by totat-hours index ) 

The basic indexes, employment and total hours worked, differ funda- 
mentally in that the hourly unit is constant while the employment unit 
ranges from part-time work during the depression years to full-time in 
the recovery. An employee working three days a week is not, for the 
purpose of the comparison shown by the derived indexes, the same 
"man" as one working five or six days a week. 

The curves in Chart 115 are not indicative of either decline or in- 
crease in the average contribution of labor. It is highly unlikely that 
efficiency per man or per hour decreased during the depression. In fact, 
if the general tendency in the reduction in working forces that took 
place, to retain as far as possible the highly skilled workers, is taken 
into account, it is probable that the average productive capacity of 
labor rose rather than fell during that period. The curves reflect in the 
earlier years the general slowing-down of the whole productive process, 
due to lack of demand and disorganized markets. In the rise that fol- 
lowed, improved equipment and economies in production methods 
developed from the experience gained during the depression contrib- 
uted heavily to the gain in rate of production. These factors alone 
accounted to a considerable extent for the increase in ratios of output 
per man and per hour. 

The foregoing examples, covering a few significant factors in a typi- 
cal economic problem, illustrate some applications of the index method 
in statistical research and analysis. The numerous economic indexes 
published by public and private statistical agencies can often be em- 
ployed in background studies for comparison with developments on a 



STATISTICAL PROCEDURES 



187 



OUTPUT PER MAN AND PER HOUR 




80 



60 



1929 



'31 '32 '33 '34 

CHART 115. 



35 



'36 



'37 



smaller scale in individual enterprises. As the indexes are immediately 
available, studies of this general nature can be made with a minimum 
expenditure of time and effort. All the basic indexes used in the index- 
number examples were taken from a single volume, the Statistical 
Abstract of the United States, published by the U.S. Department of 
Commerce. The tables were done in one day by one person, including 
computation of the derived indexes. 

If the calculations are made without the aid of computing machines, 
the closest possible rounding consistent with a fair degree of accuracy 
should be employed to shorten and simplify the work. Even with ma- 



188 GRAPHIC PRESENTATION SIMPLIFIED 

chines the use of large figures is usually unnecessary. Investment in 
a good slide rule will pay handsome dividends in economy of time and 
tedious computation. Anyone can learn to use a slide rule with an hour 
or so of practice. For computing percentages or other ratios which do 
not run beyond three digits, it is faster than a machine. All of the 
derived series in the index-number tables were computed with a twenty- 
inch slide rule. A smaller size may be used for determining the pro- 
portions of charts and for multiplication or division to two digits, but 
for close work to three digits a longer rule is required. 

Rough sketching is an invaluable aid in statistical analysis. Relation- 
ships whose significance would become apparent only after careful 
study of the same data in a table stand out instantly in a sketch. For this 
purpose pencil sketches on printed graph paper will serve quite as well 
as inked charts plotted exactly to scale. 



XVI 



PLANNING, TABULATION, AND 
PRESENTATION OF TABLES 




HE first requisite in planning a 
table is a clear understanding of what the table is for and how it is to 
be used. Once the objectives are determined and the sources outlined, 
the succeeding steps are: 

1. Lay-out of tabulating sheet 

2. Tabulation, followed by check to sources 

3. Addition of columns and computation of derived figures, if any, 
followed by check 

4. Lay-out of stubs, columns, and column heads for typing 

5. Typing, followed by check to tabulating sheet 

TABULATING SHEETS 

Careful design of the lay-out for the tabulating sheet will obviate 
complexity and confusion and save time, labor, and mistakes. It should 
be arranged so that, as far as possible, numbers to be compared are in 

189 




190 



PLANNING AND PRESENTATION OF TABLES 191 

the same column. Comparison of numbers is easier when they appear 
one above the other rather than side by side. Units should be aligned 
vertically with units, and decimals with decimals. In tabulations con- 
taining both basic and derived figures a blank column should be left at 
the right of each basic column for recording the derived figures. In the 
final typed table the basic columns should go side by side at the left of 
the table and the derived columns side by side at the right. The reader 
compares derived figures with each other, not with the basic figures. 

The series of numbers and stub entries should be broken down into 
groups for convenience in tabulation, copying, and reading. In some 
tables, of which Table XV is an example, the arrangement in groups is 
determined by the nature of the data. Otherwise the usual practice 
is to leave a blank line after each five items. Monthly figures can be 
arranged in quarterly or semiannual groups. 

Ledger-ruled paper is often used for tabulating purposes, but the 
columns are usually too wide, with the result that the table occupies 
more space than is necessary, is awkward to handle, and must be folded 
two or three times to get it to a size convenient for filing. For general 
use a specially prepared form about ll-by-17 inches in size, ruled in 
columns one-half inch wide, with quarter-inch horizontal ruling, pro- 
vides sufficient space for all but extremely wide tables and when folded 
once will fit in a standard filing cabinet. 

UNITS OF VALUE 

When the unit of -value is the same for all the columns, it is customary 
to place it under the title ( see Table XV ) . If it applies to some of the 
columns, but not to all, it may be lettered in a box head, as shown in 
Table XVI, on page 192; and, if it applies to a single column only, it 
may be shown at the foot of the column head. 

TITLES 

The title should be a concise but fairly complete description of the 
subject-matter. The facts the table shows should be emphasized by 
naming them at the beginning of the title. Such forms as "Table Show- 
ing/' "Number of," and "Distribution of in table titles and column 
headings should not be used unless they are essential to a clear descrip- 
tion of the data. 

In the typed copy, titles should be centered above the table. If they 











K> VO CMCOlf\CUCO VOCMCOlTN 






rH 


i 


M< ajg 


CVJ ^ rH^^rHr- ^rHt-^ 






-1 





1 rH OJ 

S c i t 








O 


rH 


<D O 05 <n 

< s 


O ^i ON if\ CM CO iT\ N^OMPvCM 
ON CM cO-*rHr--ct CMCO-*H 






-p 

C 


g 




CO <H h- roo VO rT\ , it rc\ O 






i 






Vb R ft2<S^rH g^R^C^ 






PH 




1 Tj 

i -H a> 


ir\ co -H/or rr\o co-3-ot^- 


CO 








^s 


^ - S^ftaS 1 ^*ft 


i 








<! -tf > 


CM ir\ rHr--H/ot^- if\rH r-^t 

rH -=f CTVO rOv ONVO ^t O VO KN, 
O K^ OMTvCMCOlfN f^NOMfNCM 


M 








K S * 


CO rH t-- H 


8 
fe 


I 









t"^ ON VO rTN. O VO K> rH t^ rTN O 


w 


1 


rH 


M 

rH 


H PH 
r-l 


KN IP\ CM ONVO CM O\ t~- rT\ O\VO 
CM -* 


>< 


1 


rH 


O 





H -4- rHCOtAHCO VO CMXX3 ITN 
O K^\ O t^-* Of lT\rH t -* 
ON CM O\VO K> ONVO -d- O VO rfN 


. 


B 


O 


1 


03 <C 
M 


CO rH CO ifN. CM CO lf\ K^OMT\CM 
VC ON 


3 


Is- 


3 




1 

H TJ 

rH 0) 


ITv CO VO fOONVO KN rHr rTNO 

-* r-- ir\CMcoir\cu ovo CM ON 

K-N VO -H; rH r--H/ rH CTMnrHCO 


i 


s? 


c 

H 






CM IT\ rT\ OVO N"N O CO -^ O C"*- 

H -* CM ONITNCVI ON f ro ONVO 

O^ KN CO rHCO VOCMCOlTS 

ON CM" 


I 


g^ 

-0 rH 
5l ^ 


I 


t 


rH 1 


CO rH rH t J" O r- IfNrH t^--^ 
VO ON ON ITN. CM CO lf\ rfNONU^CM 


^ 


PH 


I 

-P 


a 



J -H -rl 00 ,Q 

8 "<3 


IA CO CO -* rH f-3- OJCO -^ rH 


I 





P 




SH t, 



$ % SSS^ R^St 


r"^ 




w 




If 


O K"N -H/ OVO KN. ON CO-^" O VD 
ON CM KN. ON ITN. CM CO r-rrvONiT\ 


| 


co 






" 1 


t* O CM CO -^J" rH t^ VO CM CO -^ 
VO ON rHf K"NOVO u-Xr-lf-'KN 
IfN CO OVOCMONlfN -d-OVOCM 


o 

mC^ 








r 

CO 


-4- f O\lT\rHCO-* rTNONlPvrH 
r("N VO CO -4" O r rT\ CMCO -H; O 

CM^ tr\ t-N^ONVO r-lt--rrNON 

rH -d" 


g 






















. : i? 


w 













P3 






4 
t 

I 

J 

a 
C 

c 
t 


) 
1 
4 

PJ 

; ? 

r 

i 3 (-1 
> o 

i l 

i 


2 . . -rH 
BJ rH 

<a 3 

I -5 s *s|^ 

. . -( CD -0 

W a) <Df^ -Hgq 
< SZ Bt-i OO 'tt)G-H 
OS rH " PL, 4J a) H 

OT < EH CO EnO CO 


H 



192 



PLANNING AND PRESENTATION OF TABLES 193 

run into more than one line the second line also should be centered 
and, if possible, be shorter than the first. Abbreviations should not be 
used in titles and subtitles. 

STUBS 

The items in the stub should be edited to as uniform a maximum 
length as is possible without employment of unfamiliar abbreviations. 
Long stub items should be written in two or more lines. In tables where 
the list of items is broken down into groups the usual practice is to 
letter the title of the group at the left of the stub column and indent 
the subordinate items two or three spaces. This arrangement is shown 
in Table XV. If the subordinate items are longer than the group title, 
the arrangement may be reversed, as in Table XVI. 

The order of items in the stub should follow a logical sequence. 
This may be quantitative, progressive, chronological, or alphabetical. 
Ascending or descending order of rank in the attribute in which the 
items are compared facilitates easy reading and interpretation. Any 
miscellaneous or combined group, such as "All other," should appear 
at the end of the list rather than in order of its size. The total may be 
the first or the last line of the table. 

TABULATION 

The data should be recorded in black pencil. The No. 2 or 2% is 
best for this purpose. A pen should not be used, as frequent era- 
sures are unavoidable. An erasing shield will help to prevent smudgy 
erasing. 

Much time will be saved in tabulation by working in twos. One per- 
son reads off the figures from the original source, the other records 
them on the tabulating sheet. The data for a single table may be drawn 
from a number of documents, reports, schedules, or other sources, 
which differ in shape, size, and arrangement. One person working 
alone on such a tabulation will take three or four times as long as two > 
working together and will make more mistakes. The same method will 
save time in "reading back" to sources. Errors are more easily noted 
when read aloud. 

Verification by reading back, however, is not an absolute guarantee 
of accuracy. The work is monotonous, and, if it is continued for some 
time, the attention may wander because of boredom or fatigue. Where 



194 GRAPHIC PRESENTATION SIMPLIFIED 

tape adding machines are used, the tapes, if properly labeled, can 
often be used to save time and work in verification. Some tables permit 
the use of a method that reduces materially the probability of errors. 
The lines as well as the columns are totaled. If the entries and the 
additions have been made correctly, the sum of the line totals will 
exactly coincide with the sum of the column totals. 

COLUMN WIDTHS 

In typed tables the width of the columns should be as uniform as 
the size of the figures and the number of the columns will permit. If 
crowded for space, the minimum width necessary for individual col- 
umns can be determined by counting the digits and punctuation marks 
in the largest figure of each column and adding two blank spaces for 
ruling. Long column heads" should be typed in several lines to fit the 
space required by the figures below them. The last three columns in 
Table XVI are examples. The practice of basing the width of columns 
on the space needed for long column heads written in one or two lines 
often results in columns that are far wider than is necessary for the 
data and a table that is too wide for typing on a machine of stan- 
dard size. 

CAPITALIZATION 

Titles should be capitalized throughout. In typed tables spaced 
capitals underlined are often used for short titles. Only the initial let- 
ter of the first word in subtitles is capitalized, with the exception of 
words normally written with initial capitals. The same rule applies to 
stub titles and column headings. 

PUNCTUATION 

The title should not end with a period. If it consists of two sentences 
or independent phrases, a period follows the first but is omitted after 
the second. Periods are not used in column headings or stub items 
except where required for abbreviations. Commas should be used to 
separate thousands from hundreds and millions from thousands. Lead- 
ers, a line of spaced periods between stub titles and the first vertical 
ruling, are usually included in typed tables. Ditto marks should not 
be used. 



PLANNING AND PRESENTATION OF TABLES 195 

If the data represent money, the dollar sign should be typed before 
each number in the first line, and repeated before the total if this is 
placed at the bottom of the table. It may be omitted if the data are 
rounded and the rounded unit is indicated below the title or in a 
column head. 

SYMBOLS 

All spaces to the right of each stub item must be filled in. The sym- 
bols used to indicate that the data for a particular item are lacking 
vary widely. The zero, the single dash, and the double dash are often 
used for this purpose and are quite as often misunderstood. The data 
may be lacking for various reasons. The collection of information 
concerning a particular item may have been discontinued, or the 
information may not be pertinent or necessary for the purpose of 
the table, or the data may not have been received when the tabula- 
tion was made. A footnote defining or explaining the omission is 
preferable to arbitrary symbols whose meaning is often not clearly 
understood. 

The footnote symbol most commonly used in typed tables is a com- 
bination of an underlined letter and an oblique, _a/. It occupies only 
two spaces but requires backspacing for the underline. Another, less 
frequently employed, is a combination of letter and parentheses, (b). 
The use of the asterisk and double asterisk is usually limited to tables 
with not more than two footnotes. 

RULING 

Methods in use for ruling statistical tables differ to some extent in 
minor details, but the following rules are generally applicable. The 
ruling is done in India ink with a ruling pen. A light double rule or a 
single heavy rule is drawn from margin to margin above the stub and 
column heads. The horizontal separations of the box and column heads 
are in single light rules. Single light vertical rules are drawn between 
columns. The sides of the table should be left open, without ruling. 
Totals and grand totals should be separated from other numbers by 
heavier single rules, vertical if the totals are at the right, horizontal 
if they are at the bottom of the table. Averages should be set off in 
the same manner. 



INDEX 



Abbreviations for time scales, 13 
Ahesive screens, 7, 77 

on display charts, 147 

on maps, 107 

samples of, 6 

Airbrush, coloring with, 151 
"All-items" bar, 70-71 
American Standards Association, 3, 10 
Amount scales, distribution of, 12, 13 

labeling of, 14 

number of units in, 14 

100 percent base for, 18, 20-21 

position of, 14 

repetition of, 86 

rulings for, number of, 14, 16, 17 

single, 43 

zero base for, importance of, 18 

omission of, 18 
Apprentice training, 8 
Area-bar charts, 67 
Arithmetic charts, 46, 52 
Arithmetic scales, 44, 45 
Army Service Forces, Control Division, 

Statistics Branch, 50 
Averages, 177 

averaging, 179 

moving, 26-27 

progressive, 27 

simple, 177 

in tables, 195 

unweighted. See Simple. 

weighted, 178 

of wide range of values, 177 
Ay res, Leonard P., 5 

Band charts. See Strata charts. 
Bar-and-symbol charts, 65-67 



197 



Bar charts, 54-75 

advantages of, 54 

base figures in, 70-71 

broken bars in, 54-55 

compared with circle charts, 95 

compared with sector charts, 96, 97 

components in, order of, 58-59 

keys in, 63-64 

lay-out of, 133-138 

lettering on, 60 

lettering versus scales in, 55-58 

100 percent base in, 68-69 

ruling in, 59-60 

stubs in, 71-72 
Bar pictorial charts, 113-115 
Bars, broken, 54 

on map charts, 108 

optical distortion in, 67-68 

segmented, use of black in, 67 
Base figures in relative-number charts, 

70 

Base lines, weight of, 27 
Board meetings, charts for, 165 
Borders, 13, 60 
Bow pens, 144, 145 
Brinton, Willard C., 1, 147 
Business-activity charts, 40 

Capitalization, in tables, 194 
Cardboard, mounting on, 152 
Charts, area-bar, 67 

arithmetic, 46, 52 

band. See Strata charts. 

bar, 54-75, 95, 96, 97, 133-138 

bar-and-symbol, 65-67 

bar pictorial, 113-115 

business-activity, 40 



198 



INDEX 



Charts, contd. 
cardboard-mounted, 152 
circle, 90-95,123-125, 141-142 
circular-bar, 125-126 
classification of, by use, 160-174 
column, 76-89, 122, 133-136, 138-140 
components of, 10-11, 13 
condensation afforded by, 5 
for conferences, board meetings, and 

hearings, 165-168 
cumulative-curve, 28, 29 
curve, 10-31, 77-79, 133-136, 140 
for desk use, 164-165 
deviation-bar, 69-70 
deviation column, 88-89 
display, 146-152 
enlargement of, 155, 157 
for exhibits, 168-170 
frequency-distribution, 119-120 
on graph paper, 127, 128, 130, 131 
grouped-bar, 62, 63 
grouped-column, 83-84 
horizontal-bar. See Bar charts, 
index log, 52 
index-number, 18-21, 47 
for inter-office distribution, 163 
joined-bar, 61, 136-137 
line. See Curve charts, 
logarithmic. See Rate-of-change 

charts. 

on loose-leaf paper, 165, 173 
map, 104-111 
multiple-scale, 30 
multiple-time-scale, 30 
organization, 123 
paired-bar, 64-65 
pictorial, 112-115 
pie. See Sector charts, 
progress, 174 

for public addresses, 170, 172-173 
for publication, 161-163 
ranking, 116-119 
rate-of-change, 44-53 
ratio. See Rate-of-change charts, 
for recording current operations, 173- 

174 

reduction of, 155-159 
relative-variation, 120-122 
reproduction of, 153-159 
rolled, 151-152 
sector, 96-103, 141-142 
semi-logarithmic. See Rate-of-change 

charts. 



Charts, contd. 

sliding-bar, 74, 75 

spaced-bar, 61, 138 

staircase-curve, 30-31, 41 

statistical map. See Map charts. 

step-bar, 73 

stock-price, 28, 29 

strata, 33-34 

surface-curve, 32-43 

typewritten, 130, 131 

on unruled paper, 132-145 

versus tables, 5 

vertical-bar. See Column charts. 
Charts: How to Make and Use Them, I 
Circle charts, 90-95 

compared with bar charts, 94, 95 

comparison of equal amounts with, 
123-125 

errors in drafting, 90-91 

lay-out of, 141-142 

size of, method for calculating, 91-93 
Circles, on column charts, 84 

on maps, objections to, 108 
Circular-bar charts, 125-126 
Class intervals, 119, 180 
Colors, adhesive, 7 

in contrasting net changes, 38 

on display charts, 147, 149, 151 

on graph-paper charts, 130 

on map charts, 105 

on multisegmented bars, 75 

in organization charts, 123 

in printed charts, 162-163 
Column charts, 76-89 

compared with curve charts, 77-79 

with curves, 84-86 

labels in, 76-77 

lay-out of, 133-136, 138-140 

relative-variation, 122 

spaced and joined columns in, 80-82 

with tables, 82-83 

uses of, 76 
Columns, hollow, 80 

labeling, 76-77 

on map charts, 108-110 

and spacing, 139 

width of, in tables, 194 
Compass, 141, 142 

Components, chart, designation of, 10- 
11, 13 

order of, in bar charts, 58-59 
Conferences, charts for, 165-168 
Crayons, use of, on display charts, 151 



INDEX 



199 



Cumulative-curve charts, 28, 29 
Current operations, charts for record- 
ing, 173-174 
Curve charts, 10-31 

bases in, 18-21 

compared with column charts, 77-79 

curve patterns in, 17-18 

keys in, 31 

lay-out and plotting of, 133-136, 140 

omission of time-scale ruling in, 23- 
24 

plotting points in, 24 

proportions of, 13-14 

scale numbers in, 14 

scale rulings in, 14, 16, 17 

time scales in, 13, 22 
Curves, broken, 28 

on column charts, 84-86 

fitted, 22-23 

on map charts, 108-110 

moving-average, 26-27 

patterns of, 17-18 

progressive-average, 27 

steep, distortions produced by, 34 
Cut-out letters, 143 

Dash lines, for connecting segments, 77 
Desk use, charts for, 164-165 
Deviation-bar charts, 69-70 
Deviation-column charts, 88-89 
Diameters, circle, calculation of, 92 
Display charts, 146-152 

coloring of, 147, 149, 151 

equipment for, 147-148 

lay-out of, 147-148 

lettering of, 148-149 

limitations of, 146-147 

mounted on cardboard, 152 

rolled, 151 
Distortion, in bars, 67-68 

in pictorial charts, 112 

produced by steep curves, 34-35 
Dot maps, 107-108 
Drafting, inducements for learning, 7 

mathematical requirements for, 176 

order of operations in, 145 
Drafting courses, 9 
Drafting table, 147-148 
Drawing board, 132 
Drop pens, 144, 145 

Easels, for chart exhibits, 170 
Elementary Statistics, 176 



Enlargements, 155, 157 
Equipment, 127, 132 

for display charts, 147-148 

improvements in, 6-7 
Exhibits, chart, 168-170 
Explanatory notes, 48, 50 

Flow maps, 111 

Footnotes, 195 

Freehand lettering, 6-7, 143, 148-149, 

161, 166 

French curve, 132 

Frequency-distribution charts, 119-120 
Frequency distributions, 179-180 

Gelatin duplicators, 155 
General Staff conferences, 4 
Geometric forms, 91 
Graphic Methods for Presenting Busi- 
ness Statistics, 1 
Graphic Methods for Presenting Facts, 

Graphic Presentation, 1 
Graph paper, charts on, 127, 128, 130, 
131 

samples of, 129 
Grids, defined, 13 

dimension of, 13-14 

lay-out of, 133-134 

rate-of-change, ruling of, 44-45 
scale for laying out, 49 

weight of lines in, 27 
Grouped-bar charts, 62, 63 
Grouped-column charts, 83-84 

Hand-lettering, 6-7, 143, 148-149, 161, 

166 

Harper, Dr. F. H., 176 
Hatching, by hand, 7 
Hearings, charts for, 165-168 
Horizontal-bar charts. See Bar charts. 

Illustration board, use of, for display 

charts, 148 

Indexes, basic and derived, 186 
Index log chart, 52 
Index-number charts, base in, 18-21 

versus rate-of-change charts, 47 
Index numbers, 181-188 
Inking, 144 
Inter-office distribution, charts for, 163 



200 



INDEX 



Joined-bar charts, 61 
lay-out of, 136-137 

Keys, for bar charts, 63-64 
for curve charts, 31 
for grouped-column charts, 83-84 
for map charts, 111 

Lantern slides, 172-173 
Lay-out, circle-chart, 141-142 

column-chart, 138-140 

curve-chart, 140 

display-chart, 147-148 

grid, 48, 133-134 

joined-bar, 136-137 

sector-chart, 141-142 

spaced-bar, 138 
Leroy Lettering Guide, 149 
Lettering, on bars, 55-58, 60-61 

basic and derived figures, 86-87 

with cut-out letters, 143 

on deviation-bar charts, 69-70 

on display charts, 146, 148-149 

hand, 6-7, 143, 148-149, 161, 166 

irregularities in, 166 

with lettering guide, 6, 143, 148-149, 
161, 166 

points of greatest interest, 32, 38, 8.8 

printed, 162 

for printed reproduction, 161-162 

radial, 99 

of sector charts, 99 

stripped, 161-162, 164 

typewritten, 130, 143, 155, 161 

versus scales, 55-58 
Lettering-guide alphabet, 150 
Lettering guides. See Lettering. 
Lighting, for chart exhibits, 170 

for conferences, 166, 167-168 

for public addresses, 170, 172 
Line charts. See Curve charts. 
Logarithmic charts. See Rate-of-change 
charts. 

Map charts, 104-111 
with circles, 108 
coloring, 105 
with columns, bars, or curves, 108- 

109 

with dots, 107-108 
flow, 111 
keys for, 111 
limited uses of, 104 



Map charts, contd. 

outlines for, 104 

with pins or tacks, 110 

with surface designations of cate- 
gories, 105-107 
Map tacks, 110 
Median, 178 
Microphone, use of, in public addresses, 

173 
Military Committees, Senate and 

House, 5, 167-168 
Mimeoscope, 155 
Moving-average curves, 26-27 
Multilith, 155 
Multiple-scale charts, 30 
Multiple-time-scale charts, 30 

Nomenclature, standardization of, 3 
Norm, 69 

100 percent base, 18-21, 68-69 
Operations, order of, 145 
Optical distortion, in bar, 67-68 

produced by steep curve, 34-35 
Order of values, ascending, 59 

descending, 59 
Organization charts, 123 
Outline maps, 104 

Paint, use of, on display charts, 150- 

151 

Paired-bar charts, 64-65 
Paper, 143 
Parallel rule, 148 
Percents, averaging, 179 

rounding, 177 
Photo-offset, 154 
Photostat, 153-154 
Pictorial charts, 112-115 

symbols for, 113-114 
Pie charts. See Sector charts. 
Pin maps, 110 
Plotting, of amounts, bar charts, 137, 

138 

column charts, 139-140 

curve charts, 140 
points, 24-25 
yearly and monthly, 22 
Progress chart, 174 
Progressive-average curves, 27 
Protractor, 141, 142 
Public addresses, charts for, 170, 172- 

173 



INDEX 



201 



Publication, charts for, 161-163 
Punctuation, in tables, 194 

Quartermaster General, 165 

Radial lettering, 99 
Ranking charts, 116-119 
Rate-of-change charts, 44-53 

explanatory note necessary in, 48, 50 

grid ruling of, 44-46 

grids, lay-out of, 48 

limitations on use of, 50, 52-53 

multicurve, 50 

omission of scale rulings in, 46 

versus index numbers, 47 
Ratio charts. See Rate-of-change 

charts. 

Reductions, 155-159 
Relative-variation charts, 120-122 

column design, 122 

curve design, 121 
Reproduction processes, 153-159 
Rolled charts, 151-152 
Root-two dimensions, 13-14, 15 
Rough sketching, 188 
Rounding, 176-177 
Ruling, in bar charts, 58, 59, 60 

combined with ticks, 16, 17 

in curve charts, 14, 16, 17, 23-24 

elimination of, 43 

excessive, 14, 16, 17 

graph-paper, samples of, 129 

rate-of-change scale, 44-46 

replaced by ticks, 16, 17 

of tables, 195 
Ruling pen, 144 

Scale rulings. See Ruling. 

Scales, arithmetic, 44, 45 

elimination of, to gain space, 43 
for laying out rate-of-change grids, 

49 

tape, 134-136 
triangular decimal, 134 
use of, in bar charts, 55-58 
See also Amount scales, Time scales. 

Sector charts, 96-103 

compared with bar charts, 96, 97 

lay-out of, 141-142 

lettering of, 98, 99 

miniature, 100-101 

overloading of, 102 

use of shading in, 101-102 



Sectors, grouping of, 99 

order of, 96, 99 

separation of, for emphasis, 101 

totals of, indicating, 102-103 
Segment labels, bar charts, 57 

column charts, 76-77, 87 
Semi-logarithmic charts. See Rate-of- 
change charts. 

Shading, adhesive screens for, 6, 7, 77, 
107, 147 

care in using, 67-68 

contrasting net changes with, 38 

differentiation of trends with, 32, 33 

in grouped-bar charts, 63 

in grouped-column charts, 83 

in map charts, 105, 107 

in sector charts, 99, 101-102, 103 

for showing net results, 36 

for showing significant time periods, 
36-38 

in staircase-curve charts, 41 

versus solid black, 42 
Simple average, 177 
Slide rule, for computing diameters, 92 

for computing index numbers and 
percentages, 188 

in determining reductions or en- 
largements, 159 
Sliding-bar charts, 74, 75 
Spaced-bar charts, 61 

lay-out of, 138 
Square roots of decimals, 93 
Staircase-curve charts, 30-31, 41 
Statistical map charts. See Map charts. 
Statistics Branch, Control Division, 

Army Service Forces, 50 
Statistics Branch, General Staff, 3 
Stencils, 155 
Step-bar charts, 73 
Stock-price charts, 28, 29 
Strata charts, 33-34 
Stripping, 161-162 

Stubs, adjustment of, to save space, 71- 
72 

defined, 10, 13 

lettering, on bars, 60-61 

ruling under, 58 

in tables, 193 
Surface-curve charts, 32-43 

for business-activity charts, 40 

contrasting net changes in, 38 

differentiation of trends in, 32-33 



202 



INDEX 



Surface-curve charts, contd. 

miniature, 38-40 

shading versus black in, 42 

showing net results in, 36 

significant time periods in, 36-38 

staircase-curve design of, 41 

strata, 33-34 

Symbols, in bar-and-symbol charts, 65- 
67 

footnote, 11, 195 

in tables, 195 

Tables, 189-195 

capitalization in, 194 

column widths in, 194 

combined with charts, 82-83 

preparation of, 189 

punctuation in, 194-195 

ruling of, 195 

stubs in, 193 

symbols in, 195 

tabulating sheets for, 189, 191 

tabulation of, 193-194 

titles of, 191, 193 

units of value in, 191 

versus charts, 5 
Tabulation sheets, 189 191 
Tabulation, 193-194 
Tack maps, 110 
Tape scales, 134-136 
Text summaries, 165 
Ticks, combined with ruling, 16, 17 

defined, 10, 13 

ruling replaced by, 16, 17 
Time scales, abbreviations for labeling, 
13 

break in, 21-22 

distribution of, 12, 13 



Time scales, contd. 

in multiple-time-scale charts, 30 

number of rulings for, 14, 16, 17 

as plotting points, 24 

ruling of, omission in curve charts, 
23-24 

separation between years in, 14 
Time series, 10 

Time Series Charts, A Manual of De- 
sign and Construction, 1 
Titles, chart, 13 

table, 191, 193 

typewritten, 130, 131 
"Total" bar, 70-71 
Totals, in tables, 195 
Training, methods of, 8-9, 127-131 

reduction in length of, 7-8 
Trends, projection of, 10 
Triangles, 132 

use of, in lay-out and plotting, 134, 

136, 137 

Triangular decimal scale, 134 
T-square, 132 

use of, in lay-out, 134, 136 
Typewritten charts, 130, 131 

Units of value, tables, 191 
Unweighted average. See Simple 
average. 

Values, order of, 59 

units of, in tables, 191 
Vari-Typer, 143, 165 
Vertical-bar charts. See column charts. 

Weighted average, 178-179 
Zero base, 18, 46, 120 



FUNK & WAGNALLS 

New Desk 

Standard 

Dictionary 



Featuring EMPHATYPE the surest, simplest way of 
showing how to pronounce a word literally at a glance 
this big dictionary brings you everything you need in 
a desk dictionary for home or office use. 

It is new and modern in every way : 

100,000 Funk & Wagnalls-tested definitions 
make it more complete than any other dictionary 
at anywhere near the price 

It contains a full coverage of synonyms, anto- 
nyms, an appendix of 12,000 geographical items 

850 illustrations, tables, lists 

and, a special feature of this desk dictionary, A 
HANDBOOK OF SECRETARIAL PRACTICE, 
packed with practical, specific information on 
office procedures and forms, the correct spelling, 
pronunciation, grammar of any word you're 
likely to want to use, practical hints on addresses 
and salutations, business letter forms, erasures 
and corrections, carbon copies and many others. 

This is a work dictionary, especially designed for handy, 
everyday desk duty at home or in the office. 

Bound in Blue Vellum, 6y 4 x 9, 960 pages, plain $3.50; 
thumb-indexed $3.75 

FUNK & WAGNALLS COMPANY 
153 East 24th Street New York 10. N. Y. 



Qtk&i MoJl&u* 





THE LAW OF FREE ENTERPRISE 

Lee LoeVinger, Larson, Loevinge* & Lindquist; Minneapolis 

A complete, coherent account of the legal framework of our economic 
system, in plain non-technical language. Tells the origins, present scope 
and interpretation^ the significance and probable lines of future devel- 
opment of the laws which outline our economic freedom. $5.00 

SUPERVISION IN BUSINESS AND INDUSTRY 

Robert D. Loken, Director of Employee Training, Neiman-Marcus Company 
Earl P. Strong, Professor of Management, University of Illinois 

A practical handbook of duties and methods for the modern supervisor, which 
presents for management and first line supervisors the basic principles of super- 
vision in simple, easy-to-put-into-practice style. The material in it has been applied 
in thousands of work situations and the consequent improvement in the quality 
of supervision reflected in increased output and higher morale. $3.50 

THE ECONOMICS OF INDUSTRIAL MANAGEMENT 

Walter Rautenstrauch and Raymond Villers 

Consultants in Industrial Management, Members of the Faculty, 
Department ?f Industrial Engineering, Columbia University 

To help you eliminate guesswork in management, it answers such vita ['questions 
as: how can a man stay in business? how can he make a profit? In non-technical 
language, it shows how to visualize your business through Rautenstrauch's Profit 
and Loss and Break-Even Charts. It discusses the problems of debt retirement, 
depreciation, cost of output, inflation, population and expansion. It is the practical, 
professional approach to the problems of the economic management of an industrial 
enterprise. $5.00 

HOW TO REDUCE DISTRIBUTION COSTS 

Richard D. Crisp, Director of Market-Research, Tat ham-Laird, Inc. 

A positive tested program for increasing sales and reducing marketing 
, costs. Includes methods for measuring sales performance, manpower 
requirements, territorial variations in sales costs, salesmen's activities. 
Emphasizes practical aspects of sales control techniques, provides de- 
tailed guidance in solving specific sales management problems. $5.00 



UNK & WAGNALLS COMPANY NEW YORK 10